VOLATILE ARSENIC IN AQUATIC ENVIRONMENTS€¦ · troduced until the early 1970s with the invention of the hydride generation technique. Arsenic bio-methylation pathways in different

VOLATILE ARSENIC IN AQUATIC ENVIRONMENTS

Der Fakultät für Geowissenschaften, Geotechnik und Bergbau

der Technischen Universität Bergakademie Freiberg

eingereichte

DISSERTATION

zur Erlangung des akademischen Grades

doctor rerum naturalium

Dr. rer. nat.

vorgelegt

von Dipl. Geol. Britta Planer-Friedrich

geboren am 15.09.1975 in Gunzenhausen

Freiberg, den 27.08.2004

III

ABSTRACT To understand the behavior of metals and metalloids in the environment, speciation is one of the most important requirements because occurrence, sorption, mobility, potential assimilation, and toxicity depend significantly on the individual species. The release of As to groundwater in Bang-ladesh by reduction of As(V) to As(III) is a well-known example for the complexity and possible consequences of species conversions. Yet, inorganic As(III) and As(V) are not the only important As species. Trivalent and pentavalent organic (methylated) species act as connecting links for the transfer between hydrosphere and biosphere. The most reduced forms of organic and inorganic As species are volatile and, thus, enable transport from the hydrosphere to the atmosphere. They are also the most toxic As species known. Their accidental release has been documented as early as in the mid of the 19th century. Species-selective determination methods, however, have not been in-troduced until the early 1970s with the invention of the hydride generation technique. Arsenic bio-methylation pathways in different organisms have been investigated, but reports about the occur-rence of dissolved methylated and especially volatile As species in aquatic environments are still rare. Quantification has seldom been achieved and several compounds such as volatile sulfur- or chloroarsines have not even been qualitatively identified in nature, yet.

A major requirement for identifying volatile As compounds in the environment is a simple stan-dardized sampling procedure that guarantees long-term stability of the sample. In the present re-search work a robust method for in-situ sampling of volatile metallics from aqueous environments was developed. Efficient gas-water separation is achieved with a PTFE membrane (porosity 0.1 µm) inside a PTFE collector cell that can be placed at the sediment-water interface or in the aque-ous body itself. Vacuum up to -500 mbar can be applied to force the gases to a trapping system of either liquid or solid sorbents. As liquid sorbent a 1:100 diluted NaOCl solution in a PTFE bottle is used. PTFE rings in the oxidizing bottle increase the reaction surface and thus trapping efficiency that was 84% ± 6.6 for As in the laboratory. The build-up of a high pressure from the intense gas evolution during hydride generation might be responsible for the loss of some volatile As, espe-cially at the beginning of the reaction. Quantitative evaluation of volatile metallics concentrations is achieved by referring the trapped mass in the oxidizing solution [µg] to the replaced vacuum volume [m3].

Besides this fast screening method on total volatile metallics concentrations, two species-selective sampling techniques were investigated for As. Solid phase micro extraction (SPME) fibers enabled the detection of more volatile As species in lower concentrations than solid sorbent tubes. For SPME fibers, selective sorption was observed dependent on the volatile As species. Volatile chloro-arsines (CH3)AsCl2 and (CH3)2AsCl with larger molecular weights were sorbed on polydi-methylsiloxane fibers whereas polydimethylsiloxane fibers with Carboxen and divinylbenzene coating proved best for MMA, DMA, and TMA. The fibers with only Carboxen coating showed by far the largest peak areas of all fibers for DMA, TMA, and (CH3)2AsCl. The most effective sorp-tion material in sorption tubes was also Carboxen but only TMA and traces of (CH3)2AsCl were detected. Probably the selected Carboxen type (CAR 564) is not the optimum fit yet. The volatile

IV

As species were stable on the SPME fibers for at least 3 days when wrapped in aluminum foil and stored at 4°C in the refrigerator. Quantification could not be achieved mainly because of unknown sorption affinities of the individual arsines, non-availability of calibration standards, and potential competitive sorption with other, unknown gases.

As an example for an aqueous environment predestined for the release of volatile metallics Yellow-stone National Park was investigated. It is located over the largest continental hot spot world wide and contains more than 10,000 geothermal features. After the first detection of volatile As during a reconnaissance study in 2002, 5 study areas (Nymph Lake, Hazle Lake, Ragged Hills, Gibbon Gey-ser Basin, and Lower Geyser Basin) were sampled for their water and gas chemistry. Sulfate-dominated waters with low pH and As concentrations between 20 and 2,000 µg/L were differenti-ated from chloride-dominated waters with near neutral pH and As concentrations between 1,050 and 11,000 µg/L. The precipitation of As minerals was thermodynamically calculated for Hazle Lake (orpiment supersaturation) and actually observed for a low temperature hot spring at Ragged Hills (precipitation of an amorphous Fe, S, and As rich material).

Volatile As was found in all samples besides volatile Al, B, Ba, Cu, Fe, K, Li, Si, Sr, Zn, and S. By species-selective sampling on SPME fibers, four volatile As species were detected: (CH3)3As, (CH3)AsCl2, (CH3)2AsCl, and (CH3)2AsSCH3. The latter three were all found for the first time in a natural environment. Modeling of the gaseous As species proved impossible because of a lack of thermodynamic constants for the species detected. The volatile inorganic As, As-S, As-F, and As-Cl species for which thermodynamic data were available showed negligible partial pressures of <10-11 Vol%. Total concentrations in the trapping solution from passive, diffusion based sampling were converted to concentrations in gas phase using Fick´s first law. Medium concentrations were 18 mg/m3 for sulfate-dominated waters and 50 mg/m3 for the chloride-dominated ones. Concentra-tions from active gas sampling by pumping were lower (20-100 µg/m3) but are supposed to be sub-ject to significant dilution by withdrawing ambient air from over-pumping. The detected concentra-tions are significantly increased compared to the acute toxicity limit for 1 hour exposure (160 µg/m3) with respect to AsH3. However AsH3 is probably not the predominant volatile species. Even though concentrations decrease rapidly with increasing distance from a geothermal feature animals grazing and warming up at hot springs might be exposed to significant amounts of volatile metal-lics, particularly As.

Further research on volatile metallics in aquatic environments should focus on quantification of individual volatile species, on their chemical or microbial origin, their distribution, and their toxic-ity potential.

V

ZUSAMMENFASSUNG Um das Verhalten von Metallen und Metalloiden in der Umwelt zu verstehen, ist die Speziierung eine der wichtigsten Anforderungen, da Vorkommen, Mobilität, mögliche Assimilation und Toxizi-tät entscheidend von den jeweiligen Spezies abhängen. Die Freisetzung von As ins Grundwasser in Bangladesh durch Reduktion von As(V) zu As(III) ist ein gut bekanntes Beispiel für die Komplexi-tät und möglichen Konsequenzen von Speziesumwandlungen. Doch anorganisches As(III) und As(V) sind nicht die einzigen wichtigen As Spezies. Dreiwertige und fünfwertige organische (me-thylierte) Spezies stellen das Bindeglied für den Transfer zwischen Hydro- und Biosphäre dar. Die reduziertesten Formen von organischem und anorganischem As sind flüchtig und ermöglichen somit einen Transport von der Hydro- in die Atmosphäre. Sie sind auch die As Spezies mit der höchsten bekannten Toxizität. Ihre unbeabsichtigte Freisetzung wurde bereits im 19. Jahrhundert dokumentiert. Speziesselektive Bestimmungsmethoden wurden jedoch erst in den frühen 1970ern mit der Erfindung der Hydridgenerierungstechnik eingeführt. Biomethylierungspfade von Arsen in verschiedenen Organismen wurden untersucht, aber Berichte über das Vorkommen von gelösten methylierten und vor allem flüchtigen As Spezies in der Hydrosphäre sind immer noch rar. Eine Quantifizierung wurde selten erreicht und verschiedene Verbindungen, wie z.B. volatile Schwefel- und Chlorarsine, wurden bisher nicht einmal qualitativ in der Natur nachgewiesen.

Eine wesentliche Anforderung für den Nachweis von flüchtigen As Verbindungen in der Umwelt ist ein einfaches standardisiertes Verfahren, das eine Langzeitstabilität der Probe garantiert. In der vorliegenden Forschungsarbeit wurde eine robuste Methode für die in-situ Probenahme volatiler Metall(oid)e in wässrigen Medien entwickelt. Eine effiziente Gas-Wasser Trennung wird erreicht über eine PTFE Membran (Porosität = 0.1 µm) in einer PTFE Sammelzelle, die an der Grenz-schicht Wasser-Sediment oder im Wasserkörper selbst eingebaut werden kann. Ein Vakuum bis zu -500 mbar kann angelegt werden, um die Gase in eine Sorptionslösung oder auf einen Feststoffsor-benten zu überführen. Als Sorptionslösung wird eine 1:100 verdünnte NaOCl Lösung in einer PTFE Flasche verwendet. PTFE Ringe in der Oxidationsflasche erhöhen die Reaktionsfläche und somit die Sorptionseffizienz, die im Labor für As bei 84% ± 6.6 lag. Der große Druckaufbau infol-ge intensiver Gasentwicklung während der Hydridgenerierung ist vermutlich verantwortlich für einen Teil des Verlusts an volatilem As, vor allem zu Beginn der Reaktion. Eine quantitative Be-stimmung der Konzentrationen an flüchtigen Metall(oid)en wird erreicht, indem man die Masse an sorbierter Substanz in der Oxidationslösung [µg] auf das verdrängte Vakuumvolumen [m3] bezieht.

Neben dieser schnellen Screening-Methode für Gesamtkonzentrationen an volatilen Metall(oid)en, wurden zwei speziesselektive Probenahmetechniken für As untersucht. Mit Festphasenmikroex-traktion (soild phase micro extraction SPME) - Fasern waren mehr volatile As Spezies in geringe-ren Konzentrationen nachweisbar als mit Festphasen-Sorptionsröhrchen. Für SPME Fasern wurde eine selektive Sorption in Abhängigkeit von der jeweiligen volatilen As Spezies beobachtet. Wäh-rend die volatilen Chloroarsine (CH3)AsCl2 und (CH3)2AsCl mit dem höheren Molekulargewicht auf Polydimethylsiloxan-Fasern sorbiert wurden, erwiesen sich die Polydimethylsiloxan-Fasern mit Carboxen und Divinyl-Beschichtung am geeignetsten für MMA, DMA, und TMA. Die nur mit Carboxen beschichteten Fasern zeigten bei weitem die größten Peakflächen aller Fasern für DMA,

VI

TMA, und (CH3)2AsCl. Das effektivste Sorptionsmaterial in Sorptionsröhrchen war ebenfalls Car-boxen, aber nur TMA und Spuren von (CH3)2AsCl wurden nachgewiesen. Wahrscheinlich war der gewählte Carboxen-Typ (CAR 564) noch nicht optimal. Eingewickelt in Aluminium-Folie und bei 4°C im Kühlschrank gelagert waren die volatilen As Spezies auf den SPME Fasern mindestens 3 Tage stabil. Eine Quantifizierung konnte nicht erreicht werden, vor allem wegen unbekannter Sorp-tionsaffinitäten der einzelnen Arsine auf den SPME Fasern, fehlender Kalibrierstandards und mög-licher Konkurrenzreaktionen mit anderen, unbekannten Gasen um Sorptionsplätze.

Als Beispiel für ein aquatisches System, das prädestiniert für die Freisetzung von volatilen Me-tall(oid)en ist, wurde der Yellowstone National Park untersucht. Er befindet sich über dem größten kontinentalen Hot Spot weltweit und umfaßt mehr als 10.000 geothermale Strukturen. Nach dem ersten Nachweis von volatilem As während eines Erkundungstrips 2002 wurden 5 Untersuchungs-gebiete (Nymph Lake, Hazle Lake, Ragged Hills, Gibbon Geyser Basin und Lower Geyser Basin) bezüglich ihrer Wasser- und Gaschemie beprobt. Sulfat dominierte Wässer mit niedrigem pH und As Konzentrationen zwischen 20 und 2.000 µg/L konnten unterschieden werden von Chlorid do-minierten Wässern von nahezu neutralem pH-Wert und As Konzentrationen zwischen 1.050 und 11.000 µg/L. Die Ausfällung von As Mineralen wurde thermodynamisch berechnet für Hazle Lake (Orpiment-Übersättigung) und tatsächlich beobachtet für eine Niedrig-Temperatur Quelle in Rag-ged Hills (Ausfällung von amorphem Fe-, S- und As-reichen Material).

Volatiles As wurde in allen Proben neben volatilem Al, B, Ba, Cu, Fe, K, Li, Si, Sr, Zn und S ge-funden. Über die speziesselektive Probenahme mittels SPME Fasern wurden vier volatile As Spe-zies nachgewiesen: (CH3)3As, (CH3)AsCl2, (CH3)2AsCl und (CH3)2AsSCH3. Die letzten drei wur-den alle zum ersten Mal in natürlicher Umgebung gefunden. Die Modellierung volatiler As Spezies erwies sich als unmöglich aufgrund fehlender thermodynamischer Konstanten für die gefundenen Spezies. Die volatilen anorganischen As, As-S, As-F und As-Cl Spezies, für die thermodynami-schen Konstanten vorhanden waren, zeigten vernachlässigbare Partialdrücke von <10-11 Vol%. Die Gesamtkonzentrationen in der Sorptionslösung aus der passiven, diffusiven Probenahme wurden über das 1. Fick´sche Gesetz umgerechnet in Konzentrationen in der Gasphase. Mittlere Konzentra-tionen waren 18 mg/m3 in den Sulfat dominierten Wässern und 50 mg/m3 in den Chlorid dominier-ten. Die Konzentrationen aus der Gasprobenahme durch Pumpen waren geringer (20-100 µg/m3), unterliegen aber vermutlich erheblicher Verdünnung durch mitangesaugte Umgebungsluft. Die nachgewiesenen Konzentrationen sind deutlich erhöht im Vergleich zum Grenzwert akuter Toxizi-tät für AsH3 bei einstündiger Exposition (160 µg/m3). Allerdings ist AsH3 vermutlich nicht die prädominante volatile As Spezies. Obwohl die Konzentrationen mit zunehmender Entfernung von den Geothermalstrukturen rasch abnehmen, könnten Tiere, die an den heißen Quellen grasen oder sich wärmen einer erheblichen Menge an volatilen Metall(oid)en, insbesondere As, ausgesetzt sein.

Weitere Forschung über volatile Metall(oid)e in aquatischen Systemen sollte sich auf die Quantifi-zierung einzelner volatiler Spezies, ihren chemischen oder mikrobiellen Ursprung, ihre Verteilung und ihr Toxizitätspotential konzentrieren.

VII

ACKNOWLEDGEMENT Three years of study on a subject yet untouched by scientists of our department and quite „chemi-cal“ for a geologist did make up for a very interesting and instructive time. I would like to thank Prof. Dr. Jörg Matschullat from my home university Technische Universität Bergakademie Freiberg, Interdisciplinary Environmental Research Center, for his supervisorship, his support on papers and posters during this time and proof-reading of my thesis. Some lessons were hard to learn especially in the beginning spending hours in the laboratory without any apparent success. Sincere thanks for some decisive hints and the chemical discussions to Prof. Dr. Gerhard Roewer (Department of Inorganic Chemistry, Technische Universität Bergakademie Freiberg).

My second supervisor, Dr. Darrell Kirk Nordstrom (U.S. Geological Survey Boulder / USA), did not only enable first access to Yellowstone National Park and all important contacts on a spontane-ous email from an unknown German PhD student but also provided laboratory facilities and en-dured me and my master students during several stays in Yellowstone and Boulder. Special thanks for his proof-reading of this thesis and his never-ending efforts on improving my English.

The whole PhD project would never have been realized without the financial support of the Ger-man National Academic Foundation who did not only provide a 3-years-scholarship plus 2 travel grants to USA but were refreshingly non-bureaucratic towards my plans of interrupting my schol-arship for half a year to organize an international conference and accept a DAAD short term lec-tureship in Mexico. Special thanks here to Dr. Max Brocker, Dr. Angelika Wittek and Dr. Matthias Frenz, as well as to their contact persons here in Freiberg, Prof. Dr. Gert Grabbert and Prof. Dr. Karl Lohmann. Thanks also to the PhD program of the Technische Universität Bergakademie Freiberg and its coordinators Dr. Corina Dunger and Edda Paul for supporting 2 Yellowstone field trips. The Leisler-Kiep Travel Grant awarded by Dr. Walther Leisler-Kiep in November 2003 helped to continue research in USA even after all other funds were exhausted.

Development of a new equipment is always difficult but it would have been impossible without the creativity of Michael Sekul and his crew at the workshop of the Technische Universität Berga-kademie Freiberg constructing PTFE collector cells, oxidizing bottles, thermal desorption units, etc. from draft sketches. Their splendid performance and never-ending patience helped to improve the equipment up to the last detail. For laboratory assistance on the (sometimes tricky) GF-AAS I would like to thank Dipl. Chem. Peter Volke, for introduction to the GC-MS and assistance at vari-ous column changes Dipl. Chem. (FH) Hans-Joachim Peter. Thanks also to Prof. Dr. Matthias Otto, Dr. Silke Lehmann, and Dipl. Ing. (FH) Gisela Sachse from the Department of Analytical Chemis-try for access to their GC-MS.

For providing free sample material for my initial research studies I would like to thank Dr. Michael Marquardt at Varian (Chromosorb 105, 106), Eric Butrym at Scientific Instrument Services (Car-bosieve), Dr. Petra Münch at Sigma Aldrich (Carboxen, Carbotrap), Dr. Susanne Faulhaber at Sig-ma Aldrich (CAR/DVB fiber), and Maike Pieper at Infiltec (PTFE membrane sample).

VIII

An especially instructive part of my own work was supervising master students in various study areas and I´d like to acknowledge the good cooperation with Manja Seidel, Silke Mannigel, Anja Landgraf, Katja Zimmermann, Cornelius Goldhahn, Sandra Göhler, Ulrich Knauthe, and Juliane Becker.

Several research groups in Germany and abroad have provided equipment or valuable scientific advices during my research. For scientific cooperation and discussions about volatile As mass spec-tra I would like to thank Prof. Dr. Alfred V. Hirner and Jan Kösters (Department for Environmental Analytics and Applied Geochemistry, University Essen / Germany). For supervising the master thesis on volatile As in a constructed wetland thanks to Prof. Dr. Hendrik Emons and Zita Sebes-vari (at that time Research Center Jülich / Germany). Dr. Yong Cai and Sheena Szuri (Department of Chemistry, Florida International University Miami / USA) offered great hospitality and assis-tance when investigating the Florida Everglades. Prof. Dr. Johnnie Moore (University of Montana Missoula / USA) was the first to suggest volatile As studies at Yellowstone besides the original study area Milltown Reservoir in Missoula and provided his laboratory facilities. His assistant Dr. Heiko Langner was not only a valuable support for all my analyses but he and his family also of-fered me a warm and vivid home for 6 weeks.

Special thanks goes to Dr. Henry Heasler, Christie Hendrix, Shannon Savage, and Steve Miller (National Park Service, Yellowstone Nationalpark, Wyoming) for their support with Park bureauc-racy, logistics and field work especially during our trip in September 2003. The spontaneous great cooperation with Dr. Earl Mattson and Dr. Mitchell Plummer (Idaho National Engineering and Environmental Laboratory Idaho Falls / USA) during the same trip is gratefully acknowledged. Further thanks goes to Prof. Dr. Tim McDermott and Dr. Corinne Lehr (Montana State University Bozeman / USA) for organizing GC-MS access at their university and especially to Dr. John Gar-barino, Dr. Mark Sandstrom and Kevin Fehlberg (United States Geological Survey, Federal Center Denver / USA) for numerous GC-MS analyses in a fascinating laboratory environment.

My acknowledgements to the colleagues Dr. James W. Ball and Blaine R. McCleskey (United States Geological Survey Boulder / USA) do not only include assistance in the field and especially during laboratory work, but extend to the pleasure it is working within their research group. Fi-nally, Yellowstone field work would have never been as enjoyable as it was without Manja Seidel, Beate Böhme and Juliane Becker accompanying me in June/July and September/October 2003, without Betty DeWeese´s warm hospitality (Yellowstone River Motel), and of course without sur-vey, organizational, and scientific help from Steve Dill (Federal Highway, Northwestern Division).

Sincere thanks to my parents and my sister Kerstin for their interest and patience in sharing the ups and downs during my research work. For his scientific support, for believing in me, and for keep-ing my ambitions high by always putting new tasks ahead I would like to thank Broder.

For the benefit and enjoyment of the people...

Roosevelt Arch, North Entrance to Yellowstone National Park (ca. 1929) YNP Photo Archives

XI

LIST OF CONTENTS

Introduction ........................................................................................................................... 1

1 Methylated and volatile metallics....................................................................................... 5

2 Arsenic.............................................................................................................................. 11

2.1 Geogenic occurrences ............................................................................................................. 11 2.2 Anthropogenic use .................................................................................................................. 12 2.3 General toxicity....................................................................................................................... 13 2.4 Global balances ....................................................................................................................... 14 2.5 Species .................................................................................................................................... 15

2.5.1 Dissolved inorganic arsenic .............................................................................................. 17 2.5.1.1 Inorganic As(V) .......................................................................................................... 17 2.5.1.2 Inorganic As(III) ......................................................................................................... 18

2.5.2 Dissolved organic arsenic species ..................................................................................... 22 2.5.2.1 Mono- and dimethylated As(V) acids ......................................................................... 26 2.5.2.2 Tri- and tetramethylated As(V) acids.......................................................................... 28 2.5.2.3 Mono- and dimethylated As(III) acids........................................................................ 29 2.5.2.4 Arsenosugars, arsenocholine, arsenobetaine............................................................... 31 2.5.2.5 Organic arsenic sulfur compounds.............................................................................. 32 2.5.2.6 Organic arsenic halogen compounds .......................................................................... 33

2.5.3 Volatile arsenic species ..................................................................................................... 35 2.5.3.1 Inorganic AsH3............................................................................................................ 35 2.5.3.2 Monomethylarsine ...................................................................................................... 36 2.5.3.3 Dimethylarsine ............................................................................................................ 36 2.5.3.4 Trimethylarsine ........................................................................................................... 37 2.5.3.5 Organic Sulfur Arsines................................................................................................ 38 2.5.3.6 Inorganic and organic halogen arsines........................................................................ 38

2.6 Analytical speciation techniques............................................................................................. 39 2.6.1 Arsenic separation............................................................................................................. 39

2.6.1.1 Hydride generation...................................................................................................... 39 2.6.1.2 HPLC .......................................................................................................................... 43 2.6.1.3 Gas chromatography ................................................................................................... 44 2.6.1.4 Capillary zone electrophoresis .................................................................................... 44

2.6.2 Arsenic detection............................................................................................................... 45 2.6.3 Arsenic quantification ....................................................................................................... 46

2.7 Field sampling......................................................................................................................... 47 2.7.1 Water sample stability and preservation ........................................................................... 47 2.7.2 On-site species separation for water samples.................................................................... 49

XII

2.7.3 Gas sampling without pre-concentration........................................................................... 50 2.7.4 Gas sampling with pre-concentration................................................................................ 51

2.7.4.1 Liquid sorbents............................................................................................................ 51 2.7.4.2 Solid sorbents .............................................................................................................. 51 2.7.4.3 Cryotrapping ............................................................................................................... 55

2.7.5 Techniques for collecting volatile metallics...................................................................... 56 2.7.5.1 Collection of volatile metallics from a gas phase ....................................................... 56 2.7.5.2 Collection of volatile metallics dissolved in an aqueous phase .................................. 57

2.8 Summary of chapter 2 ............................................................................................................. 58

3 Development of a field method........................................................................................ 59

3.1 Development of an in-situ gas-collector cell........................................................................... 59 3.2 Gas generation and trapping.................................................................................................... 61

3.2.1 Optimization of arsine generation ..................................................................................... 62 3.2.2 Optimization of screening on total concentrations (liquid sorbents)................................. 65 3.2.3 Species-selective sampling (solid sorbents) ...................................................................... 70

3.2.3.1 Sorption tubes ............................................................................................................. 70 3.2.3.2 Solid-phase micro-extraction (SPME) ........................................................................ 73

3.3 Summary of chapter 3 ............................................................................................................. 77

4 Yellowstone National Park............................................................................................... 79

4.1 Introduction ............................................................................................................................. 79 4.1.1 Geology of the Park .......................................................................................................... 80

4.1.1.1 Pre-Quaternary Geology ............................................................................................. 80 4.1.1.2 Quaternary Volcanism ................................................................................................ 81 4.1.1.3 Quaternary Glaciation ................................................................................................. 83 4.1.1.4 Hotspot development after the last volcanic cycle...................................................... 84 4.1.1.5 Hydrogeochemistry of geothermal features ................................................................ 86

4.1.2 Study areas ........................................................................................................................ 89 4.1.2.1 Nymph Lake area ........................................................................................................ 91 4.1.2.2 Hazle Lake area........................................................................................................... 93 4.1.2.3 Norris Geyser Basin .................................................................................................... 94 4.1.2.4 Gibbon Geyser Basin .................................................................................................. 98 4.1.2.5 Lower Geyser Basin.................................................................................................... 99

4.2 Methodology ......................................................................................................................... 100 4.2.1 Field methods .................................................................................................................. 100

4.2.1.1 Reconnaissance trip 2002.......................................................................................... 100 4.2.1.2 Sampling trips 2003 .................................................................................................. 102

4.2.2 Laboratory ....................................................................................................................... 106 4.2.3 Data processing ............................................................................................................... 107

XIII

4.2.3.1 Water data check....................................................................................................... 107 4.2.3.2 Gas concentration conversion ................................................................................... 110 4.2.3.3 Gas chromatography interpretation........................................................................... 115 4.2.3.4 Statistics .................................................................................................................... 115 4.2.3.5 Speciation modeling.................................................................................................. 115

4.3 Results ................................................................................................................................... 116 4.3.1 Cluster analysis ............................................................................................................... 116

4.3.1.1 Clustering of the water samples ................................................................................ 116 4.3.1.2 Clustering of the gas samples.................................................................................... 119

4.3.2 Description of water and gas chemistry .......................................................................... 120 4.3.2.1 Type I steam-heated waters....................................................................................... 120 4.3.2.2 Type II deep thermal waters...................................................................................... 130

4.3.3 Modeling As species and mineral phases........................................................................ 138 4.3.4 Volatile As speciation ..................................................................................................... 141 4.3.5 Modeling volatile As speciation...................................................................................... 145 4.3.6 Quantification of volatile metallics ................................................................................. 146

4.4 Summary of chapter 4 ........................................................................................................... 151

5 Recommendations .......................................................................................................... 153

6 References ...................................................................................................................... 155

7 Appendix ........................................................................................................................ 173

XIV

LIST OF ABBREVIATIONS a activity AAS atomic absorption spectrometry AES atomic emission spectrometry AFS atomic fluorescence spectrometry AMD acid mine drainage AMP adenosine monophosphate amu atomic mass unit Apoptosis normal series of events in a cell that lead to its death (and then replace-

ment), apoptosis can also be induced by external influences, cancer cells avoid apoptosis

Ars arsenic resistance ArsR protein that encodes a repressor protein ArsA protein that transports an energizing subunit, produces As(III)-specific

ATPase ArsB protein that translocates the As(III) via a transmembrane efflux channel

driven by the ATP hydrolysis produced by ArsA ArsC arsenate reductase; protein that converts As(V) in the cell to As(III) which

is pumped out of the cell via the ArsA/ArsB anion pump ArsD regulatory protein As(V) inorganic pentavalent arsenic (H3AsO4

0, H2AsO4-, HAsO4

2-, AsO43-)

As(III) inorganic trivalent arsenic (H3AsO30, H2AsO3

-, HAsO32-, AsO3

3-) ATP adenosine triphosphate ATPase protein complex responsible for converting electrical potential energy into

ATP b.p. boiling point cAMP cyclic ester of AMP CAR Carboxen cytotoxicity toxicity to cells CW CarbowaxTM CZE capillary zone electrophoresis DMA gaseous dimethylarsine, (CH3)2AsH DMAA organic dimethylated arsenic acid, without distinction between tri-or pen-

tavalent species DMAVA organic pentavalent dimethylated arsenic acid [(CH3)2AsO(OH),

(CH3)2AsO2-]

DMAIIIA organic trivalent dimethylated arsenic acid [(CH3)2As(OH), (CH3)2AsO-] DNA deoxyribonucleic acid DTT dithiothreitol, C4H10O2S2 DVB divinylbenzene EDTA ethylenediaminetetraacetic acid ESI electrospray ionization ET-AAS electrothermal AAS fd film thickness (GC column) FI-HG flow injection HG GC gas chromatography GF-AAS graphite furnace AAS GFP green fluorescent protein (GS)2AsSe seleno-bis(S-glutathionyl) arsinium ion GSH glutathione HG hydride generation HG-AAS hydride generation AAS HPLC high performance liquid chromatography

XV

HRTEM high resolution transmission electron microscope ICP inductively coupled plasma ID inner diameter (GC column) L length (GC column) mesh size largest diameter of sorption material grains, e.g., 60 mesh = grains that

passed through 0.25 mm mesh size, 80 mesh = 0.18 mm, 100 mesh = 0.15 mm, 200 mesh = 0.07 mm

Michaelis constant measure for the enzyme-substrate affinity, the higher the Michaelis con-stant, the weaker the affinity

MMA gaseous monomethylarsine, (CH3)AsH2 MMAA organic monomethylated arsenic acid, without distinction between tri-or

pentavalent species MMAVA organic pentavalent monomethylated arsenic acid [(CH3)AsO(OH)2

0, (CH3)2AsO2(OH)-, (CH3)AsO3

2-] MMAIIIA organic trivalent monomethylated arsenic acid [(CH3)As(OH)2,

(CH3)2AsO(OH)-, (CH3)2AsO22-]

m.p. melting point MS mass spectrometry MS/MS tandem mass spectrometry Operon functional unit of the DNA which synthesizes a special protein PA polyacrylat PDMS polydimethylsiloxane PE polyethylene PFA perfluoroalkoxy, teflon, m.p. 310°C, tensile strength 110 kg/cm2 at 250°C Pit non-specific and fast phosphate inorganic transport system, active at high

phosphate concentration Pst energy requiring, specific phosphate inducible transport system, works in

times of phosphate depletion PTFE polytetrafluorethylene, teflon, m.p. 327°C, tensile strength 85 kg/cm2 at

250°C REE rare earth elements (La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb,

Lu) Repressor protein the product of a regulatory gene that blocks the function of another gene RSH thiols, monovalent -SH radicals attached to a carbon atom, mercaptan RS-SR disulfide SAHC S-adenosylhomocysteine SAM S-adenosylmethionine SEM scanning electron microscope SPE solid-phase extraction SPME solid-phase micro extraction TETRA tetramethyl arsonium ion (CH3)4As+

Thiols monovalent -SH radicals attached to a carbon atom, mercaptan TMA gaseous trimethylarsine, (CH3)3As TMAO trimethylarsineoxide, (CH3)3AsO TPR templated resin transferase enzyme that realizes the transport of groups from their donor to their ac-

ceptor

INTRODUCTION

On January 22, 2001, the U.S. Environmental Protection Agency (EPA) lowered the enforceable maximum contamination level for As from 0.05 mg/L to 0.01 mg/L finally following the provi-sional guideline advised in the World Health Organization guidelines for drinking-water quality (World Health Organization 1993) and enforceable drinking-water standards in other countries (e.g., TrinkwV 1990). The revision was preceded by various discussions on the level to be pro-posed including 0.003 mg/L as minimum feasible level and 0.005 mg/L as proposed on June 22, 2000. Numerous disputes and contradictions caused a further delay of the effective date until Feb-ruary 22, 2002 to conduct additional health effect, risk, cost, and benefit reviews. Public water sys-tems must comply with the 0.01 mg/L standard beginning January 23, 2006 (Environmental Protec-tion Agency 2001).

This uproar in the As community is a result of new knowledge about As toxicity and its global occurrence in increased concentrations both naturally and man-made. Since the late 1980s, natu-rally high As concentrations in West Bengal and Bangladesh have become known as the cause of „the worst mass poisoning of a population in history“ (Smith et al. 2000). Groundwater wells drilled to supply the population with clean drinking-water without microbial contamination pene-trated peat layers in the sedimentary basin which caused reduction of iron oxyhydroxides and re-lease of their sorbed arsenic load to solution (Ravenscroft et al. 2001). Examples of generally more localized, man-made As contaminations include sites of mining activities, metal processing, high-temperature combustion, and electronic and chemical industries (pesticides, growth stimulants, pharmaceuticals, glass ware, galvanizing, tanning).

High total As concentrations are, however, not the only concern: Occurrence, sorption, mobility, potential assimilation, and toxicity depend significantly on the individual As species. Dissolved tri- and pentavalent inorganic and organic (methylated) as well as volatile As species can be distin-guished. A thorough review on As speciation literature revealed that within the last three decades, since species-selective determination methods were introduced, significant progress has been made on inorganic As speciation and on in vivo and in vitro studies of organic As species. Hydrogeo-chemical studies on the occurrence of dissolved methylated and especially volatile As species in aquatic environments, however, are still rare. The qualitative and quantitative detection of methy-lated and volatile compounds is important because they act as connecting links for the transfer be-tween hydrosphere - biosphere and hydrosphere - atmosphere, respectively. Furthermore, trivalent methylated and volatile As are the most toxic As compounds known.

Considering the sampling of volatile As, one major problem is extracting gases from an aqueous phase. The other problem is the stabilization of volatile As after collection because it is easily hy-drolyzed and oxidized. To solve these problems was one of the primary goals of the present re-search work. A robust technique for in-situ separation of gaseous compounds from the aquatic phase by a PTFE membrane in a specially designed PTFE collector cell was developed. The cell can be placed at the sediment-water interface or in the aqueous body and gases are withdrawn from

2 Introduction

the aqueous phase by vacuum up to -500 mbar and trapped in either liquid or solid sorbent. Trap-ping the gases in a liquid sorbent will not only allow screening of the main target element volatile As but of all other volatile metals and metalloids present at the sampling site, summarized in the following as volatile metallics.

After the initial development of the method, various study sites were chosen for a first screening on the occurrence of volatile As in solution. Study sites comprised wetlands at two abandoned ura-nium tailings in Lengenfeld and Schneckenstein / Germany, the US superfund site Milltown, Mon-tana / USA, mine water irrigated rice fields in Luojiaqiao town / China, a constructed wetland in Jülich / Germany, a smouldering heap of As-rich coal in Oelsnitz / Germany, the Florida Ever-glades / USA, and Yellowstone National Park / USA.

The screening of the different aquatic environments mainly served for optimizing the developed sampling and trapping method for field application and for selecting one model area for further detailed studies on correlation of aqueous and gaseous As chemistry and volatile As speciation. Only the results from Yellowstone National Park are presented here. A brief introduction to the other study sites is given in Appendix 1. Yellowstone National Park was selected as final study area because of the high dissolved (up to 11 mg/L) and volatile As concentrations detected here under a wide range of hydrogeochemical conditions (pH from 1.9 to 8.7, temperature from 20 to 92°C, redox potential from -260 mV to 720 mV, and anion predominance from chloride and sulfate to hydrogencarbonate). The sampling campaigns were conducted in June/July and Septem-ber/October 2003. The results are a central part of this research project besides the As review and the development of the field sampling technique for volatile metallics.

The thesis starts with a brief introduction to methylated and volatile metallics in general, their chemical characteristics, occurrence in the environment and typical concentration ranges (chapter 1). The second chapter provides a thorough review of the main target element As. Many reviews exist about As and central aspects are often global balances, toxicity, or, especially since the last 10 years, microbiology. The major focus of this review is the presentation of data available for the individual species especially the methylated and volatile species that are often not considered. Natural occurrences, chemical properties, toxicity, and abiotic and microbially catalyzed species conversion are described (section 2.5). Analytical speciation techniques with As separation, detec-tion and quantification are summarized in section 2.6. Field sampling methods are discussed in section 2.7. For aqueous As species, this section includes water sample stability and preservation as well as possibilities of on-site species separation. For gaseous As species, sampling methods with-out pre-concentration are described compared to pre-concentration on solid sorbents, in liquid sor-bents and in cryotraps.

Even though volatile As species at Yellowstone were mainly sampled directly from the gas phase, not by separation from the aquatic phase, the equipment for achieving the gas-water separation in aquatic environments is described in chapter 3. The outline of the field method development further includes the results of trapping efficiencies of various liquid sorbents (GF-AAS analysis) and tests

Introduction 3

on species-selective determinations by solid sorbents and solid phase micro extraction fibers with GC-MS analysis.

After an introduction to geology, hydrogeology and hydrogeochemistry of the main study area Yellowstone National Park, chapter 4 gives the details of field and laboratory methodology and data processing (section 4.2). Results are discussed in clusters of similar water and gas chemistry and speciation modeling is provided (section 4.3.1 to 4.3.3). For volatile As, speciation by GC-MS is described (section 4.3.4) and an attempt of hydrogeochemical modeling is presented (section 4.3.5). In the last section ( 4.3.6), an evaluation of the quantification of volatile metallics is given. One page summaries at the end of the chapters 2, 3, and 4 provide the most important information of each chapter.

A digital MS Access database included on a CD together with this thesis contains all presented information from pictures and descriptions of the study sites, raw analytical data, to data for quality control, and interpreted results.

1 METHYLATED AND VOLATILE METALLICS

Within the last 40 years, lessons learned from numerous accidents and an increasing public interest in environmental problems have led to intensive studies about the fate of metallics in the environ-ment. Interdisciplinary work between geoscientists, chemists, toxicologists and physicians has shown that it is not sufficient to know total metallics concentrations for an understanding of the complex chemistry of metallics from geogenic and anthropogenic sources in natural systems, nor are total metallics concentrations sufficient for making predictions and giving advice for remedia-tion measures. Speciation has become one of the most pressing subjects in research, since occur-rence, mobility, sorption, potential assimilation, and toxicity vary significantly between different species of the same element.

Considerable attention has been paid to the detection of inorganic metal and metalloid species, yet the occurrence of organometallics in the environment is still widely unknown. The transition proc-ess from inorganic to organometallics is called alkylation, i.e. the replacement of an OH--group by an alkyl-group. In natural processes, the alkyl group mostly is a methyl (CH3

-) group, more seldom an ethyl (C2H5

-), phenyl (C6H5-) or butyl (C4H9

-) group. Methylation is enzymatically catalyzed and occurs in a wide range of conditions from anoxic-methanic to oxic - not only in microorganisms, but also in algae, plants, animals, and humans. Few animals are known that are not capable of bio-methylation, e.g., several monkeys, chimpanzees, and the guinea pig (Healy et al. 1997; Wildfang et al. 1997). High total concentrations of organometallics are not a prerequisite. Methylation occurs at low concentrations. Purely abiotic methylation is possible, e.g., in the presence of methyliodide, but is rarely observed. Transalkylation, i.e. the transfer of an alkyl group, can occur. Jewett et al. (1975) state the following sequence for transmethylation: Pb > Sn > Hg > As. Thus, Sn can transfer a methyl group to Hg according to the reaction Sn(CH3)3

+ + Hg2+ ↔ Sn(CH3)22+ + Hg(CH3)+, while

Hg can not transfer the methyl group to Sn.

Organometallics are among the most mobile environmental compounds, soluble in polar and non-polar media, accumulating in fatty tissue, capable of passing biological membranes, and most of them readily volatile due to their high vapor pressures. Compared to their inorganic predecessors some of them are less toxic, others are more toxic. Volatile organometallics have, comparable to inorganic volatiles (hydrides), an extremely high biotoxicity potential. All compounds with a boil-ing point <150°C and a vapor pressure >103Pa (Clemens and Lewis 1988) are defined as “volatile”. The first methylated compounds detected in the beginning of the 19th century were the volatile ones. In 1824, “bismuth breath” was described as a garlic odor in the breath of mine workers who were exposed to bismuth carbonate. The same phenomenon is reported from patients treated with bismuth carbonate for stomach disorders. Later, this volatile compound was identified to be (CH3)2Te from Te, an impurity in bismuth carbonates that is biotransformed in humans (Bentley and Chasteen 2002b). After several incidents of arsenic poisoning in Germany and Great Britain, Gmelin (1839) noted that a garlic smell was present in the rooms, where the poisoning had oc-curred. This compound was identified as (CH3)3As, volatilized from wallpapers painted with As pigments (Challenger et al. 1933). Hofmeister (1894) reported strong garlic odor in the expired

6 Methylated and volatile metallics

breath of a dog after subcutaneous administration of 0.06 g sodium tellurate; another early hint on biomethylation of inorganic Te to (CH3)2Te.

Methylated mercury achieved worldwide attention. In 1956, fishermen of the small town of Mina-mata, Japan began to show symptoms of a neurological disorder. About 3 years later, the causes were traced back to methylmercury poisoning. A local company, Chisso Corporation, used Hg in the process of manufacturing acetaldehyde and dumped their processing wastes into the Minamata Bay. From 1932 to 1968, an estimated 27 tons of inorganic Hg were dumped into the bay. Some of the Hg was biomethylated to methylmercury chloride, accumulated in fish, and finally, since fish and shellfish were the main sources of food for the local residents, in humans (Takizawa 1979). Because of the “Minamata” disease Hg became the model for the formation of organometallics and many papers report its presence in the environment.

A more recent example is the methylation of Sb that became a news topic in the mid 1990s. Richardson (1994) claimed that the sudden infant death syndrome (SIDS) might be related to me-thylation and volatilization of toxic gases from flame retardants in mattresses and covers. The flame retardants contained Sb and increased Sb concentrations in the liver of SIDS victims were thought to result from trimethylstibine inhalation. Later investigations, however, found no causal relationship between SIDS and levels of trimethylstibine (Jenkins et al. 2000).

Both Hg and Sb biomethylation leads to increasing toxicity for humans. However, biomethylation may also act as a detoxification mechanism for humans. One example is Se. Frankenberger and Karlson (1989) suggested methylation and subsequent volatilization as a passive rehabilitation measure for the Kesterson Reservoir US Superfund Site, California. By the beginning of the 1980s, sediment significantly enriched in Se that was transported by the San Luis drain had accumulated in the reservoir. Evaporation further increased Se concentrations in the sediment. Bioaccumulation in flora and fauna led to toxic levels of Se among the high-end consumers of the food chain, caus-ing deaths and deformities, especially in bird species.

To date, methylated volatile and non-volatile compounds have been found in many environments such as sea water, estuaries, geothermal areas, waste heaps, sewage and digester gases, or urban air ( Table 1). Higher alkylated species (e.g., tetraethyllead, butyltin, phenylmercury, ethyldimethylar-sine, triethylarsine) are usually attributed to anthropogenic sources (Feldmann and Hirner 1995; Hirner et al. 1998). Evidence for naturally occurring ethylmercury in soils and sediments from the Everglades in Florida was presented by Cai et al. (1997). Transition metals rarely form volatile species, even though carbonyl-compounds are known for Ni from anthropogenic sources like ciga-rette smoke, carbon monoxide containing industrial hydrogen or nitrogen gas or automobile ex-haust. Feldmann and Cullen (1997) detected 0.2-0.3 µg/m3 volatile Mo(CO)6 and 0.005-0.01 µg/m3 volatile W(CO)6 in waste deposit gases. Considerably more alkylated compounds have been cre-ated in experiments with incubated microbial cultures over soil, sediment, and water, as well as from biological samples. Moreover innumerable organometallics have been synthesized chemically

Methylated and volatile metallics 7

in the laboratory. Several reviews exist about organometallics in the environment (Craig 1986; Crompton 1998; Hirner and Emons 2004; Thayer 1995).

Table 1 Environmental occurrence of volatile and dissolved inorganic and organometallics

Atom Species Occurrence As AsH3 (g) waste heap gas1; digester gas2; hot spring gas, British Columbia3

(CH3)AsH2 (g) waste heap gas1; digester gas2; hot spring gas, British Columbia3; geothermal water, New Zealand 0.08-1.9 µg/kg3; soil from domestic waste heap 205-8540 ng/kg4; soil

below gas station 1780-6720 ng/kg4; soil below coal mining industry 200-1350 ng/kg4; river and harbor sediment 0.6-21.7 ng/kg5

(CH3)2AsH (g) waste heap gas1; digester gas2; hot spring gas, British Columbia3; geothermal water, New Zealand 0.02-8.9 µg/kg3; soil from domestic waste heap 215-2320 ng/kg4; soil below gas station 1030-6620 ng/kg4; coal mining 260-720 ng/kg4; river and harbor

sediment 0.8-7.3 ng/kg5 (CH3)3As (g) waste heap gas1; digester gas2; hot spring gas, British Columbia3; geothermal water,

New Zealand 0.18-0.25 µg/kg3; soil from domestic waste heap 55-1020 ng/kg4; soil below gas station 50-250 ng/kg4; soil below coal mining industry 14-215 ng/kg4;

river and harbor sediment 0.6-11.4 ng/kg5 (CH3)2As(C2H5) (g) river and harbor sediment 7.2 ng/kg5 (CH3)As(C2H5)2 (g) river and harbor sediment 3.8 ng/kg5 As(C2H5)3 (g) river and harbor sediment 1.2 ng/kg5 As (g) (no speciation) 0.1-1.8 ng/mL gas at a cattle dipping vat site, Florida25

Bi BiH3 (g) landfill and sewage gas (up to 25 µg/m3)6 (CH3)BiH2 (g) river and harbor sediment 0.1 ng/kg5 (CH3)3Bi (g) digester gas2; landfill, sewage gas (25 µg/m3)26; domestic waste heap soil 40 ng/kg4

Cd (CH3)2Cd (g) waste heap gas1; digester gas2 Ge (CH3)GeH3 (g) geotherm. water, New Zealand 0.02-0.06 µg/kg3; coal mining ind. soil 12-50 ng/kg4

(CH3)2GeH2 (g) geotherm. water, New Zealand 0.02-0.03 µg/kg3; domest. waste heap soil 350 ng/kg4 (CH3)3GeH (g) geotherm. water, New Zealand 0.04-0.17 µg/kg3; domest. waste heap soil 430 ng/kg4

Hg (CH3)Hg+ (aq) geothermal water, New Zealand 0.1-15.3 µg/kg3; peatland pore water in Experimen-tal Lake Area, Ontario, Canada (up to 6 ng/L, average 0.25-0.65 ng/L)7; Everglades sediment (< 0.1-5 ng/g dry weight = < 0.2-2% of total Hg)8; Everglades 20% of total Hg9; high altitude lakes, Western US (0.01-0.73 ng/L = 0.02-12.3 % of total Hg)10; estuaries (78-393 fmol/L Gironde, 77-445 fmol/L Rhine, 44-1147 fmol/L Scheldt)11

(CH3)2Hg (g) digester gas2; sewage gas12; urban air13; river and harbor sediment 2.8 ng/kg5; Gi-ronde estuary 0-12 fmol/L11

Pb (CH3)4Pb (g) digester gas2; urban air, Bordeaux (0.09-95.2 ng/m3)14; estuaries (0-4.4 fmol/L Gi-ronde, 0-1.2 fmol/L Rhine, 0-8.9 Scheldt)11

(CH3)3(C2H5)Pb (g) digester gas2; estuaries (0-0.7 fmol/L Gironde, 0-0.7 fmol/L Rhine, 0-2.1 Scheldt)11 (CH3)2(C2H5)2Pb (g) digester gas2; estuaries (0-0.5 fmol/L Gironde, 0-90.5 fmol/L Rhine, 0-3.0 Scheldt)11 (CH3)(C2H5)3Pb (g) digester gas2; estuaries (0-1.7 fmol/L Gironde, 0-3.6 Scheldt) (C2H5)4Pb (g) digester gas2; estuaries (0-3.0 fmol/L Gironde, 0-0.9 fmol/L Rhine, 0-7.6 Scheldt)

Sb SbH3 (g) landfill and sewage gas (thermophilic step 6.7-14.7 µg/m3)6 (CH3)SbH2 (g) geothermal water, New Zealand 0.007-0.07 µg/kg3; soil from domestic waste heap

120-430 ng/kg4; soil below gas station 260-300 ng/kg4; soil below coal mining indus-try 110-195 ng/kg4; river and harbor sediment 0.2-9.8 ng/kg5

CH3SbO(OH)2 (aq) Florida Ochlockonee river 92 ppt, ~Bay 2.3 ppb15; river and seawater (1-12 ng/L)16 (CH3)2SbH (g) geothermal water, New Zealand 0.01-0.08 µg/kg3; soil from domestic waste heap 90-

350 ng/kg4; soil below gas station 60-120 ng/kg4; soil below coal mining industry 6-25 ng/kg4; river and harbor sediment 0.1-1.2 ng/kg5

(CH3)2SbO(OH) (aq) Ochlockonee Bay 110 ppt, Gulf of Mexico 588 ppt15; river, seawater (1-12 ng/L)16

8 Methylated and volatile metallics

Atom Species Occurrence Sb(C2H5)3 river and harbor sediment 0.1 ng/kg5

Se (CH3)SeH Mediterranean Sea 0.3-0.8 pmol/m3 in air; 60-158 pmol/m3 in ocean water17 (CH3)2Se waste heap gas1; urban air, Bordeaux (0.7-8.7 ng/m3)14; hot spring gas, British Co-

lumbia3; air over Gironde estuary17; Mediterranean Sea 0.3-1.2 pmol/m3 in air, 60-636 pmol/m3 in ocean water17; sewage sludge18; soil from domestic waste heap 140 ng/kg4; soil below coal mining industry 25-45 ng/kg4; estuaries (294-1107 fmol/L Gironde, 284-8784 fmol/L Rhine, 191-12793 fmoL/L Scheldt)11; river and harbor

sediment 1.0 ng/kg5 (CH3)2Se2 waste heap gas1; Mediterranean Sea 0.3-0.6 pmol/m3 in air

71-259 pmol/m3 in ocean water17; estuaries (25-90 fmol/L Gironde, 0-289 fmol/L Rhine, 0-1625 fmoL/L Scheldt)11; river and harbor sediment 2.1 ng/kg5

Se (g) (no speciation) ponds in Kesterson Reservoir over barley and salt grass Si (CH3)SiOH (g) waste deposit gas 3.4-17.5 mg/m3; waste composting gas 0.3-3.5 mg/m3 20 Sn SnH4 (g) waste heap gas1; harbor and estuarine sediment21

CH3SnH3 (g) soil from domestic waste heap 40-18.000 ng/kg4; soil below gas station 40-80 ng/kg4; soil below coal mining industry 2440 ng/kg4; river, harbor sediment 0.4-12.3

ng/kg5 (CH3)2SnH2 (g) soil from domestic waste heap 20-3350 ng/kg4; soil below gas station 25-55 ng/kg4;

river and harbor sediment 0.3-2.9 ng/kg5 (CH3)3SnH (g) soil from domestic waste heap 15-3100 ng/kg4; soil below gas station 3-15 ng/kg4;

soil below coal mining industry 3-9 ng/kg4; river, harbor sediment 0.1-0.5 ng/kg5 (CH3)4Sn (g) waste heap gas1; digester gas2; landfill gas, Vancouver22; soil below gas station 9-35

ng/kg4; estuaries (0-4.3 fmol/kg Gironde, 0-2.5 fmol/L Rhine, 1-30.8 fmol/L Scheldt)11; harbor and estuarine sediment23

(CH3)3(C4H9)Sn (g) estuaries (0-0.5 fmol/L Gironde, 0-1.2 fmol/L Rhine, 0-5.5 Scheldt)11 (CH3)2(C4H9)2Sn (g) estuaries (0-3.8 fmol/L Gironde, 0-3.6 fmol/L Rhine, 7.4-331.4 Scheldt)11 (CH3)(C4H9)3Sn (g) estuaries (0-72.8 fmol/L Gironde, 0-55.1 fmol/L Rhine, 0-1457.5 Scheldt)11 Sn(C4H9)H3 (g) river and harbor sediment 0.4-9.6 ng/kg5 Sn(C4H9)(CH3)H2 (g) river and harbor sediment 0.1 ng/kg5 Sn(C4H9)(CH3)2H (g) river and harbor sediment 4.5 ng/kg5 Sn(C4H9)2H2 (g) river and harbor sediment 8.6 ng/kg5

Te (CH3)2Te (g) waste heap gas1; thermophilic digester gas2; river and harbor sediment 0.2 ng/kg5 Tl (CH3)2Tl+ (aq) Atlantic Ocean <0.4 – 3.2 ng/L (3-48% of total Tl)24

* list is exemplary and does not claim to be complete, only compounds detected in the environment, no laboratory ex-periments are included; (g) = gaseous species in ambient air or gas, and dissolved species volatilized from water or sedi-ment, (aq) = aqueous species in water or sediment; reference numbers cited: 1 Feldmann 1995, 2 Feldmann and Kleinmann 1997, 3 Hirner et al. 1998, 4 Hirner et al. 2000, 5 Krupp et al. 1996, 6 Feldmann and Hirner 1995, 7 Branfireun et al. 1996, 8 Gilmour et al. 1998, 9 Cai et al. 1999, 10 Krabbenhoft et al. 2002, 11 Amouroux et al. 1998, 12 Bzezinska et al. 1983, 13 Schroeder 1982, 14 Pécheyran et al. 1998, 15 Andreae et al. 1981, 16 Andreae and Froehlich 1984, 17 Amouroux and Donard 1996, 18 Reamer and Zoller 1980, 19 Biggar and Jayaweera 1993, 20 Grümping et al. 1998, 21 Donard and Weber 1988, 22 Maillefer et al. 2003, 23 Jantzen 1992, 24 Schedlbauer and Heumann 1999, 25 Thomas and Rhue 1997, 26 Feldmann et al. 1999a

Strictly speaking, all alkylated compounds are thermodynamically unstable with respect to oxida-tion, UV irradiation, thermal influence, hydrolysis, reaction with radicals, and demethylation. Be-cause their reaction kinetics are usually slow, they can be detected under natural conditions. Mean dissociation energies can be taken to estimate the stability of organometallics, because the weak bond between the metal or metalloid and the C atom is easily replaced by much stronger bonds between C-O or C-X (X = halogens). The increasing polarity between metal and C atom as well as the presence of other ligands, like humic or fulvic acids, can further destabilize the compounds. Figure 1 shows estimates for the stability of (CH3)nM species.

Methylated and volatile metallics 9

Figure 1 Estimations for the stability of (CH3)nM species; stability increases with increasing mean disso-ciation energy, and with decreasing molar enthalpy (M = metal or metalloid; modified after Feldmann 1995)

Trimethylbismuth and tetramethyllead decompose spontaneously in the gaseous phase, whereas dissociation reactions for all compounds with higher mean dissociation energies than Sn-C-compounds are endothermal and consequently more stable ( Figure 1). Higher alkylated species are generally considered less stable than methylated species. If low-energy orbitals are available, H atoms from the organic substitute can be transferred to the metal or metalloid, leading to the forma-tion of an alkene (Davidson et al. 1976).

One of the most investigated elements, especially with regard to its organic speciation, is the metal-loid arsenic. Natural occurrences, anthropogenic use, toxicity and a global As balance will be pre-sented in the following sections, before the more common inorganic and organic As species will be discussed. The As biomethylation pathway will be described in detail with relevant chemicals and enzymes, and differences in mobility and toxicity of the respective species will be given.

Stability of (CH3)nM increases

low mean dissociation energy high

Hg-Bi-Cd-Pb-In-Zn----Sn-Sb-As-Se-Ge-Ga-P-Al-Si-C exothermal reactions endothermal reactions

high molar enthalpy low

Bi-In-Pb-Cd-Hg-Zn-Sb-As-Sn-Ga-Ge-Al-P-B-C-Si

2 ARSENIC

For an introduction to the element arsenic (As), first discovered in 1250 by the German scholastic Albertus Magnus, several detailed reviews exist, e.g., Cullen and Reimer (1989); Fowler (1983); Korte and Fernando (1991); Mandal and Suzuki (2002); Matschullat (2000); Merkel and Sperling (1998); Nriagu (1994a); Nriagu (1994b); Rüde (1996); and Smedley and Kinniburgh (2002).

2.1 Geogenic occurrences

The semi-metal or metalloid As ranks 44th in abundance in the earth´s crust with average concen-trations of 2.0 ppm in the upper crust and 1.3 ppm in the lower crust (Wedepohl 1995). It is mainly associated with magmatic or sedimentary rocks, especially iron ores ( Table 2, detailed review also in Smedley and Kinniburgh 2002). Table 2 Arsenic concentrations in different rock types (Woolson 1983)

Rock Minimum [mg/kg] Maximum [mg/kg] Mean [mg/kg] Andesite, Dacite 0.5 5.8 2.0 Basalts, Gabbros 0.06 113 2.0

Ultrabasites 0.3 16 3.0 Granite 0.2 13.8 1.5

Carbonate 0.1 20 1.7 Sandstone 0.6 120 2.0

Coal 0 2000 13 Shale, Clay 0.3 490 14.5 Phosphorite 0.4 188 22.6

Sedimentary iron ores 1 2900 400

Average concentrations in soils vary significantly, depending on the parent rock and inhomogeni-ties. Values cited are between 0.1 for sandy soils and 40 mg/kg for clayey soils or soils rich in or-ganic matter. The world-wide average is about 5-6 mg/kg (Woolson 1983). Extreme concentrations were found over Dartmoor granites in Great Britain with 250 mg/kg, quartzites in Brisbane, Austra-lia, with 100-200 mg/kg, and gold mines in Zimbabwe with 2000 mg/kg (Peterson et al. 1981).

About 245 As minerals are known (Onishi 1969). Besides the four elemental As modifications there are 27 arsenides (e.g., loellingite FeAs2, rammelsbergite NiAs2), 13 sulfides (e.g., arsenopy-rite Fe(As,S)2, cobaltite CoAsS, gersdorfite NiAsS, enargite Cu3AsS4, realgar As4S4, orpiment As4S6), 65 sulfo-salts, 2 oxides, 11 arsenites, 116 arsenates (e.g., mimetesite Pb5[Cl[AsO4)3, scoro-dite FeAsO4⋅2H2O], and 7 silicates.

High-pH groundwaters in closed basins in arid to semi-arid areas and strongly reducing groundwa-ters in organic-rich sediments with slow flushing rates tend to have higher As concentrations (> 1 µg/L) As concentrations. This pH and redox dependence is explained in the chemistry of dissolved, inorganic As (section 2.5.1). Areas known for high As concentrations in groundwater are parts of Argentina, Chile, China, Hungary, Mexico, Western USA, and with about 40 million people ex-posed to high-As groundwater, Bangladesh, Vietnam, and West Bengal. On a more localized scale, high As concentrations are also encountered in mineralized, and mined areas as well as in geother-mal areas ( Table 3). Smedley and Kinniburgh (2002) and Mandal and Suzuki (2002) provide a

12 Arsenic

review of typical As concentrations in natural waters, as well as in different countries world-wide. Also Nordstrom (2002) lists global As contaminations in groundwater.

Table 3 Examples for increased As concentrations in some aquatic systems worldwide

Country, geological setting As [µg/L] reference Groundwater, geogenic As source Argentina (loess, rhyolitic volcanic ash, thermal springs, oxidizing,

neutral to high pH, often saline) <1 to 5,300 Sancha and Castro (2001); Smedley et al. (2002)

Bangladesh (alluvial, deltaic sediments, strongly reducing, slow groundwater flow rates) <1 to 2,5000

British Geological Survey Department of Public

Health Engineering (2001) Chile, Antofagasta (Quaternary volcanic sediments, oxidizing,

arid, high salinity) 100-1000 Sancha and Castro (2001)

China, Xinjiang (alluvial sediments, reducing) 40-750 Wang and Huang (1994) Germany (Keuper sandstone) <10 to 150 Heinrichs and Udluft (1999) Hungary (alluvial sediments, organics, reducing, some groundwa-

ters high in humic acid) <2-176 Gurzau and Gurzau (2001)

Mexico, Lagunera (volcanic sediments, oxidizing, neutral to high pH) 8-620 Del Razo et al. (1990)

Taiwan (sediments, including black shales, strongly reducing, some groundwaters containing humic acid) 10 to 1,820 Kuo (1968);

Tseng et al. (1968) USA, Southern Carson Desert, Nevada (holocene aeolian, alluvial,

lacustrine sediments, reducing, high salinity) up to 2,600 Welch and Lico (1998)

Vietnam, Red River delta (alluvial, deltaic sediments, reducing, high Fe, Mn, NH4)

1-3050 Berg et al. (2001)

West Bengal (alluvial, deltaic sediments) < 10 to 3,200 Sun et al. (2001) Geothermal* New Zealand, Waimangu, Waitapu, Tokaanu 11-6,700 Hirner et al. (1998) Philippines, Mt Apo 3,100-6,200 Webster (1999) USA, Imperial Valley, California up to 15,000 Welch et al. (1988) USA, Lassen Volcanic NP, California up to 27,000 Thompson et al. (1985) USA, Steamboat Springs, Nevada up to 2,700 Welch et al. (1988)

USA, Yellowstone NP, Wyoming <1-5,800 Ball et al. (1998a, 1998b, 2001, 2002); Stauffer and

Thompson (1984) Mining areas Brazil, gold mining, Morro Velho 7,300 Williams (2001) Canada, British Columbia <0.2 – 556 Azcue et al. (1994) Mexico, Zimapan (mine tailings, but also naturally occurring ar-

senopyrite, scorodite, tennantite) < 14 – 1,000 Armienta et al. (1997, 2001)

Thailand, Ron Phibun 5,114 Williams (2001) USA, Fairbanks, Alaska (gold mine tailings, schist, alluvium) up to 10,000 Welch et al. (1988) Zimbabwe, Globe and Phoenix 7,400 Williams (2001) Zimbabwe, Iron Duke 72,000 Williams (2001)

* more examples in Webster & Nordstrom (2003)

2.2 Anthropogenic use

Arsenic-bearing minerals were already mined by the early Chinese, Indian, Greek, and Egyptian civilizations (Azcue and Nriagu 1994), and arsenic found wide-spread applications. As long ago as 55 B.C., Nero poisoned Britannicus to secure his Roman throne and inorganic As2O3 became very popular as a poison especially in the Middle Ages. Famous poisoners were the Borgias, Pope Alex-ander VI and his son Cesare Borgia. The popularity of the "inheritance powder" came to an abrupt end when in 1836 the development of Marsh´s assay created a possibility to prove arsenic in the bodies. The first poisoner to be convicted was Marie Lafarge in 1842 for the murder of her husband

General toxicity 13

(Gorby 1994). Later on, As found a widespread application in agriculture. Inorganic As2O3 was used as pesticide, Pb- and Ca-arsenates as insecticides, and organic arsenicals as herbicides. In 1974, the application (not the production!) of As pesticides was banned in Germany (Bundesminis-ter für Ernährung, Landwirtschaft und Forsten 1974) and the USA. Dimethylarsenic acid achieved notoriety as defoliant (“agent blue”) during the war in Vietnam (Léonard 1991). Poison on the one hand, remedy on the other, As was also used as a prominent medicine to treat rheumatism, malaria, sleeping sickness, and syphilis until the introduction of Penicillin in 1909 (Gorby 1994). More re-cent applications range from metal processing to chemical industries.

In 1989, As production was 53,000 t As2O3 (Ishiguro 1992). By 1993, it had decreased to 30,453 t (Matschullat 2000) and thereafter it stayed more or less constant with approximately 35,500 t in 2001, and 35,000 t in 2002 and 2003 (Brooks 2004). However, As is still released to the environ-ment from numerous sources, e.g., high-temperature combustion (oil, coal, waste, cement), com-post and dung (growth stimulant e.g., for poultry), glass ware production (decoloring agent), elec-tronic industries (GaAs or InAs as semiconductor material), ore production and processing, metal treatment (lead and copper alloys, admixture in bronze production), galvanizing industry, tanning industry (depilation agent), ammunition factories, chemical industry (dyes and colors, wood pre-servatives, pesticides, pyrotechniques, drying agent for cotton, oil, and solvent recycling), and pharmaceutical and cosmetic industry (Ishiguro 1992; Léonard 1991).

2.3 General toxicity

Arsenic is still the most common source of acute inorganic metal poisoning and is second only to lead in chronic ingestion. Symptoms of acute As poisoning are restlessness, throat discomfort, nau-sea, chest pain, vomiting, diarrhea, convulsions, collapse, and cardiac failure, sometimes even lead-ing to coma or death. Chronical effects include bronchitis, myocardial infarction, arterial thicken-ing, peripheral neuropathy, hyperkeratosis, hyperpigmentation, the so-called “black foot” disease (necrosis, mainly on palms and soles, first identified in Taiwan), skin (Col et al. 1999; Tsuruta et al. 1998), lung, bladder (Hopenhayn-Rich et al. 1996a), liver and kidney cancer, as well as teratogenic effects (inorganic As can cross the placenta), mutagenic changes, and genotoxicity (Carson et al. 1986; Florea et al. 2004; Mandal and Suzuki 2002). It is, however, often hard to eliminate other pathological factors, like malnutrition, or concurrent exposures. A detailed recent review about As toxicology is provided in Chou et al. (2000).

Main sources of arsenic uptake are food, inhalation, ingestion of dust or soil, and drinking-water. Average daily intake rates are between 12 and 50 µg. In areas where fish and shellfish, known to yield relatively high levels of arsenic, are the main diet, intake may be up to several 100 µg/day (Bentley and Chasteen 2002a). Much higher doses were consumed by people from Styria, Austria in the 19th century, for “plumpness to the figure, cleanness and softness to the skin, beauty and freshness to the complexion, as well as to improve breathing” (Wanklyn 1901). These so-called “arsenic-eaters” began with less than 30 mg doses 2 or 3 times a week and increased the dosage to

14 Arsenic

about 130 mg. In some cases, the practice continued for 40 years without apparent detrimental ef-fects. In one recorded case, a woodcutter was observed by a physician to eat about 300 mg of pure As2O3 for two days in a row without apparent health effects, whereas in many cases 65 mg of As2O3 is already fatal (Roscoe and Schorlemmer 1911).

Dermal sorption is still controversial. In experiments with rhesus monkeys, Wester et al. (1993) found a low degree of As sorption with only 2-6% of the arsenic applied finally absorbed by the skin. Arsenic uptake in plants occurs in the oxidized form as arsenate in competition with phos-phate for the same uptake carriers in the root plasmalemma. Mycorrhizal fungi are likely to play an important role in that assimilation because they enhance phosphate acquisition for the host plants (Meharg and Hartley-Whitaker 2002).

Once incorporated, As undergoes a complex pathway of chemically and enzymatically induced reactions, partly resulting in metabolites that are even more toxic than the educts. Arsenic toxicity strongly depends on the respective As species. The most toxic form of As are gaseous As com-pounds, followed by dissolved organic trivalent, inorganic trivalent, inorganic pentavalent, and organic pentavalent As compounds, and last but not least elemental arsenic and arsenosugars. Each species´ toxicity, species conversions and human detoxification mechanisms will be discussed in detail in section 2.5.

2.4 Global balances

Trying to quantify the geogenic and anthropogenic As fluxes and its sources and sinks, is extremely difficult, because many of the published quantitative data are questionable. Two major problems are sampling or analytical errors and generalizations from only locally representative results to global budgets or even a lack of precise information on the reservoirs for which fluxes are to be balanced. Several approaches exist, e.g., Lantzy and MacKenzie (1979), MacKenzie et al. (1979), Walsh et al. (1979), and Chilvers and Peterson (1987). Figure 2 shows a scheme of the As fluxes, reservoir sizes, and average background concentrations according to one of the most recent detailed studies (Matschullat 2000).

One of the most critical points in all global As budget models is quantifying the emissions from the lithosphere and the hydrosphere to the atmosphere. Based upon an assumed high release of As by volatilization, especially from the ocean, MacKenzie et al. (1979) estimated the gaseous emissions to be 300,000 t/a. As Cullen and Reimer (1989) point out, those results are probably incorrect, since they were based upon assumptions of As concentrations in rain water that were too high. Chilvers and Peterson (1987) estimate 160 to 26,200 t/a for low temperature volatilization from soils and 22,000 t/a for volcanic emissions (terrestrial and submarine). The calculated ratio of an-thropogenic to natural emissions according to them is 40:60. Matschullat (2000), however, states that especially the volcanic emissions are probably underestimated.

Species 15

Figure 2 Global As cycle with fluxes and reservoir sizes for atmosphere, biosphere, pedosphere, hydro-sphere, anthroposphere and lithosphere (modified after Matschullat 2000)

2.5 Species

In aqueous solutions As forms no single cations, but reacts readily with C, O, S, halogens or H. Predominant oxidation states are 3+ and 5+. Coordinate numbers are mainly 3 (AsF3) and 5 (AsF5), but also 4 (As(CH3)4

+) or 6 (AsF6-). Although As-C, As-O, As-S or As-halogen bonds are polarized

(Asδ+ - Xδ-), As-H bonds are not because the electronegativities of As and H are equal (EN = 2.20, Mortimer 1987). In most papers, however, As in a As-H bond is referred to as the more electro-negative element (oxidation state 3-), in others as the more electropositive (oxidation state 3+, Cul-len and Reimer 1989).

Beginning with the first positive identification of As by Marsh (1836), there was no shortage of analytical procedures for determinations of total As. However it was not until 1973 that As speci-ation became an issue, when Braman and Foreback (1973) introduced the first hydride generation technique, capable of separating different inorganic and methylarsenic compounds even in small

Atmosphere 800-1,740 t As (680-1,480 t As in the northern hemisphere, 120-260 t As in the southern hemisphere)

concentrations contributed from terrestrial sources: from oceanic sources: 0.5 - 2.8 ng/m3 (northern hemisphere) 0.072 - 0.16 ng/m3 (northern hemisphere)1.0 ng/m3 (southern hemisphere) 0.018 ng/m3 (southern hemisphere)

Hydrosphere oceans

2.38 - 5.18 ⋅ 109 t As

dissolved 2.33 ⋅ 109 t As

particulate

19.9 ⋅ 106 t As

0.1 - 1.7 µg/L

Biosphere 184,000 t As

terrestrial plants 0.5 mg/kg

Pedosphere 0.6 - 1.7 ⋅ 109 t As

5 - 7.5 mg/kg

Anthro-posphere

Hydrosphere rivers

54,000 - 61,000 t As

dissolved 3,600 - 61,000 t/a

particulate

720 - 12,240 t/a

0.1 - 1.7 µg/L

Lithosphere 4.01⋅ 1013 t As

upper crust 1.5 - 2 mg/kg, bulk crust 1 - 1.8 mg/kg

22,200 - 73,600 t/a

4,300 - 8,200 t/a

27 t/a sea spray

17,150 t/a volcanoes

30,453 t/a As-production

1,980 t/a wind erosion 160 - 26,200 t/a volatilization

25,450 t/a emissions

125 t/a natural forest fires 3,345 t/a wood, fuel fires

4,870 t/a submarine volcanism

? t/a weathering

125,000 t/a discharge

28,400 t/a residues

46,400 t/a sedimentation 38,200 t/a subduction

16 Arsenic

concentrations. Even though mobility, degradability, and toxicity vary significantly between the different species, there is still a lack of standard procedures for As speciation, especially for the methylated or volatile As species.

The As species outlined in Figure 3 will be described in more detail in the following sections. After a description of chemical properties, behavior in aqueous solutions, toxicity and transformation of the predominant dissolved inorganic species As(V) and As(III) (section 2.5.1), methylation of inor-ganic to organic As compounds and their respective properties, occurrence, and toxicity will be explained (section 2.5.2). Section 2.5.3 is dedicated to As compounds with a boiling point < 150°C, the so-called volatile arsenicals or arsines which form as intermediates and end products of the As biotransformation pathway or in a purely chemical reaction. Further details for the chemical gen-eration of hydrides are provided in section 2.6.1.1.

organic compounds

inorg. 1 CH3 2 CH3 3 CH3 4 CH3

As-Cl (III) volatile

AsCl3 CH3AsCl2 (CH3)2AsCl

As-S (III) volatile

CH3As(SR)2

(CH3)2As(SR)

As (III) volatile

AsH3

CH3AsH2

(CH3)2AsH

(CH3)3As

As (III) dissolved

H3AsO3

CH3As(OH)2

(CH3)2As(OH)

As-S (III) dissolved

CH3As(SR)2(OH)2

(CH3)2As(SR)2(OH)

(CH3)3As(SR)2

As (V) dissolved

H3AsO4

CH3AsO(OH)2

(CH3)2AsO(OH)

(CH3)3AsO

(CH3)4As+

(CH3)2AsOCH2COOH (CH3)3As+CH2COO-

arsenosugars (link to lipophils)

(CH3)2AsOCH2CH2OH (CH3)3As+CH2CH2OH

Figure 3 Selective scheme of the more common inorganic and organic As compounds, As-S-compounds (formed in the presence of H2S or thiols), and As-Cl-compounds (formed in the presence of Cl), gray line = main reaction pathway for As biomethylation as explained in section 2.5.2; shaded = volatile species

Species 17

2.5.1 Dissolved inorganic arsenic

2.5.1.1 Inorganic As(V)

The oxidized form of dissolved As is As(V). Figure 4 shows the As(V) species as a function of pH. In natural groundwaters with a near neutral pH, the negatively charged complexes H2AsO4

- and HAsO4

2- predominate. Significant amounts of the uncharged complex H3AsO40 occur in strong acid

solutions (pH < 2) only. Solution properties of H3AsO40 closely resemble those of phosphoric acid

H3PO40. The pK values for individual As species vary sometimes significantly between different

references. Therefore, the most commonly cited pK values are listed in a table next to the predomi-nance diagram. A review on thermodynamic data for inorganic As is given in Nordstrom (2000) and Nordstrom and Archer (2003).

H3AsO4

0 ↔ H2AsO4- + H+ ↔

HAsO4

2- + 2H+ ↔ AsO43- + 3H+

Figure 4 pH-dependent dissociation of As(V) species (a(As) = 1.33⋅10-7 mol/L; * data used for the diagram)

In surface water, arsenate shows a significant correlation with biological activity. Depletion in the aqueous solution during phytoplankton bloom due to uptake alternates with an increase in dis-solved arsenate in the absence of biological activity, e.g., during winter (Anderson and Bruland 1991; Cabon and Cabon 2000; Howard et al. 1995; Sanders 1985; Sohrin et al. 1997). In oxidizing groundwaters As(V) is the predominant species, e.g., in Holocene aquifers in Argentina (Smedley et al. 2002), in aquifers of Quaternary volcanogenic sediments in Chile (Thornton and Farago 1997), Mexico (Del Razo et al. 1990), or in alluvial aquifers in the Basin and Range province, USA (Robertson 1989).

Sorption on Fe-, Mn-, Al-oxides and hydroxides, clay minerals and organic matter (Chiu and Her-ing 2000; Gulens et al. 1979) as well as substitution of arsenate for SO4

2- in jarosite are probably the most important processes for As(V) removal from aqueous solutions (Savage et al. 2000; Figure 5). Smedley and Kinniburgh (2002) present a review of several solid-solution partition coef-ficients for a range of oxides and clays. At higher pH (pH > 8.5) sorption is inhibited or even de-sorption occurs. Consequently, groundwaters in semiarid or arid basins with high pH are often en-riched in As. Reductive dissolution of Fe and Mn oxides leads to As release under reducing aquifer conditions (Smedley and Kinniburgh 2002). High concentrations of phosphate, bicarbonate, sili-cate, or organic matter can also decrease As sorption due to competition for the same binding sites.

pK1 pK2 pK3 reference 2.22 6.98 11.52 Naumov et al. (1974)* 2.24 6.94 11.5 Sadiq et al. (1983) 2.25 6.94 11.59 Brookins (1988)

pK1 pK2

pK3 H3AsO40

H2AsO4-

AsO43-

HAsO42-

18 Arsenic

Alkaline earth metal-arsenates, or rather their hydrates, may form limiting mineral phases (e.g., Ca-As-hydrate; Bothe and Brown 1999). However, due to inconsistencies in the experimental setup, the initial pH, the acids used for pH adjustment, the ionic strengths, the As concentrations, the con-sideration of formation of additional mineral phases, etc., existing thermodynamical solubility products, as presented e.g., in Wagemann 1978, are often dubious. A typical example is barium arsenate (Ba3(AsO4)2 = 3 Ba2+ + 2 AsO4

3-). Current databases still list Ba3(AsO4)2 as a highly stable phase (log K = -50.110; Chukhlantsev 1956), even though Robins (1985; log K = -16.58) and Es-sington (1988; log K = -21.62) got much higher solubility products from their experiments and proved that Chukhlantsev erroneously interpreted the precipitating mineral phase BaHAsO4·H2O as Ba3(AsO4)2.

Figure 5 Predominance diagram for As(III) and As(V) species, super-imposed the predominance diagram for the Fe-S-K-CO2-system at 25°C, 1 bar, shaded fields = limiting mineral phases (modified after Savage et al. 2000)

2.5.1.2 Inorganic As(III)

Reduction of As(V) to As(III) and oxidation of As(III) to As(V) can be both enzymatically and abiotically catalyzed. For certain prokaryotes and eukaryotes As(V) can serve as respiratory oxi-dant (Ehrlich 2002). Energy is conserved if the reduction of As(V) to As(III) is coupled with the oxidation of organic matter (e.g. lactate) because the As(V)/As(III) oxidation/reduction potential is +135 mV (Oremland and Stolz 2003). Examples of dissimilatory arsenate-reducing prokaryotes (DARPs) are Pseudomonas fluorescens (Myers et al. 1973), Anabaena oscillaroides (Freeman 1985), Desulfotomaculum auripigmentum (Newman et al. 1997), Desulfitobacterium GBFH (Nig-gemyer et al. 2001), the Proteobacterium Sulfurospirillum barnesii SES-3, and Sulfurospirillum arsenophilium MIT-13 (Ahmann et al. 1997). The latter has just been recently classified as belong-

Species 19

ing to the genus Sulfurspirillum (Stolz et al. 1999). Yeast (wine yeast; Crecelius 1977a) as well as aquatic organisms like coral species (Pilson 1974) and the freshwater phytoplankton Chlorella (Blasco et al. 1972), can catalyze the reduction from As(V) to As(III).

In contrast to the DARPs several other bacteria, cyanobacteria, archaea, and eukarya (review in Mukhopadhyay et al. 2002) reduce As(V) to As(III) only as a mechanism of As detoxification and resistance not for respiration. The most well studied detoxification mechanism is the so-called “ars operon” (Oremland and Stolz 2003). The operon is a functional unit of the DNA that causes resis-tance to trivalent and pentavalent arsenicals. ArsR encodes a repressor protein that further produces 4 proteins in gram-negative bacteria: ArsA, ArsB, ArsC, ArsD. ArsA, the transport energizing sub-unit, produces an As(III)-specific ATPase, i.e. a protein complex responsible for converting electri-cal potential energy into ATP (adenosintriphosphate), the so-called “molecular motor”. ArsB trans-locates the As(III) via a transmembrane efflux channel driven by the ATP hydrolysis produced by ArsA. ArsC, the so-called arsenate reductase (Aposhian 1997; Wildfang et al. 1998) converts As(V) in the cell to As(III) which is pumped out of the cell via the ArsA/ArsB anion pump, and ArsD is a further regulatory protein. In gram-positive bacteria ArsA and ArsD are absent (reviewed in Roberto et al. 2002).

Dissimilatory iron-reducing bacteria (DIRB) and sulfate-reducing bacteria (SRB) can not only re-duce As enzymatically, but also by producing inorganic compounds (e.g., for SRB: gluthathione GSH or other thiols that are monovalent -SH radicals attached to a carbon atom RSH) that in turn reduce As(V) to As(III) abiotically (Harrington et al. 1998; Spliethoff et al. 1995). However, the purely chemical conversion rates are low. Although Cherry et al. (1979) suggested significant As(V) reduction in the presence of reduced sulfur (H2S or S2O3), especially at low pH, Harrington et al. (1998) showed that dissolved sulfide concentrations of more than 100 µM/L would be neces-sary to reduce only 10% of a 10 mM/L As(V) solution over a 30-day-period, while biotic reduction was observed to be 2 to 4-fold higher.

“Spontaneous” oxidation back to As(V) was first detected in cattle-dipping fluids in 1909 (Brün-nich 1909). A correlation between oxidation and bacterial growth was proven in 1918 by isolation of the bacterium Bacillus arsenoxydans, probably an Achromobacter (Green 1918). This bacterium, however, was eventually lost. In 1943, new species of arsenite-oxidizing bacteria were character-ized from the genus of Pseudomonas, Xanthomonas, and Achromobacter (Turner 1949). Those bacteria grow under aerobic conditions in a pH range of 6.1-9.4. Some of them, e.g., Pseudomonas acidovorans-arsenoxydans YE56, can also survive under anaerobic conditions and use nitrate to oxidize various carbon compounds, but can not oxidize arsenite under these conditions (Phillips and Taylor 1976).

Arsenite oxidizers can be both heterotrophic or chemolithoautotrophic. Heterotrophic oxidation catalyzed by a periplasmic enzyme (Mo-containing hydroxylase) is primarily a detoxification mechanism to convert As(III) to As(V) that is less likely to enter the cell (Ehrlich 2002). Chemo-lithoautotrophic oxidizer couple the oxidation of As(III) to the reduction of O2 or NO3. The energy

20 Arsenic

gained is converted to fix CO2 in organic cellular material and achieve growth (Oremland and Stolz 2003). The α-Proteobacterium NT-26 is the most rapidly growing chemolithoautotrophic arsenite oxidizer known, with a doubling time of 7.6 hours in a minimal medium of arsenite, oxygen, and carbon-dioxide-bicarbonate (Santini et al. 2000). In general, however, microbial detoxification is the more common As(III) oxidation mechanism compared to chemoautotrophism (Nordstrom 2003).

Inorganic As(III) oxidation is generally slow but can be increased by heat (Tingle 1911), ultrasonic irradiation (Witekowa and Farbotko 1972), the presence of particulate matter (Oscarson et al. 1981), metal catalysts such as MnO2, and strong oxidizing agents such as Fe(III) (Cherry et al. 1979) especially when catalyzed by light (Johnson and Pilson 1975; McCleskey et al. 2004; Nord-strom 2003; Reay and Asher 1977). Illumination with 90 W/m2 is sufficient for photochemical oxidation of 90% of a 500 µg/L As(III) solution in 2-3 h, in the presence of 0.06-5 mg/L Fe(II,III), or even more effective, 50 µM citrate as photocatalysts (Hug et al. 2001).

Oxidation rates are described by Scudlark and Johnson (1982) to be 0.3 nmol/L/day for a 0.45 ppb standard in a purely chemical reaction, 1.5 nmol/L/day considering the presence of bacteria (Pseu-domonas arsenoxydans). Wilkie and Hering (1998) describe rapid As(III) oxidation (half-life 0.3 h) in the geothermal waters of Hot Creek, California, in the presence of a group of (unidentified) bac-teria attached to submerged macrophytes; no oxidation occurred in samples after sterile filtration. Agrobacterium albertimagni (strain AOL15), later identified by Salmassi et al. (2002) as an As(III) oxidizing bacteria, might have been the unidentified bacteria in the study of Wilkie and Hering (1998). Nordstrom (2003) report a microbial oxidation rate of 3 mg/L/min at 40-60°C and pH 2.6 for a cooling hot spring flow at Yellowstone National Park. Gihring et al. (2001) identified the bacterium Thermus aquaticus in the hot springs of Yellowstone to oxidize As(III) at a rate of about 0.5 mg/L/min. Laboratory experiments indicate similar oxidation rates for Thermus aquaticus and Thermus thermophilus of about 0.14 mg/L/min after slow oxidation during the first 16 hours. Abiotic oxidation was about 100-fold less (Gihring et al. 2001). Data from McCleskey et al. (2004) indicate an even lower abiotic oxidation rate of 10-5 µmol/L/h. That is about 5 orders of magnitude lower than the biotic oxidation rate. Hambsch et al. (1995) found no As(III) oxidation in sterile samples over 14 days at all. They also suggest that there is a temperature threshold for biotic con-version, with no detected As(III) oxidation over a period of 14 days below 4°C.

Figure 6 shows the As(III) species dependence on pH. In aqueous solutions with a pH < 9 the un-charged H3AsO3

0 complex predominates. It is 4 to 10 times more soluble and mobile than As(V) (Cullen and Reimer 1989). Gulens et al. (1979) found that under oxidizing conditions at a slightly acid pH (5.6), As(III) moved 5-6 times faster through a sand column than As(V). At neutral pH, As(V) movement increased, but was still slower than that of As(III). Under reducing conditions at pH 8.3, As(III) and As(V) passed the column rapidly. While H3AsO4

0 solution properties closely resemble those of H3PO4

0, H3AsO30 rather resembles boric acid H3BO3. Unlike H3PO3 that has P-O

and P-H bonds, H3AsO30 has no As-H bonds.

Species 21

H3AsO3

0 ↔ H2AsO3- + H+ ↔

HAsO3

2- + 2H+ ↔ AsO33- + 3H+

Figure 6 pH-dependent dissociation of As(III) species (a(As) = 1.33⋅10-7 mol/L; * data used for the diagram)

In contrast to predictions derived from pe-pH diagrams about the clear predominance of As(V) in natural waters (Andreae 1979 calculated ratios of 1015-1026 AsV / AsIII for seawater), significant amounts of As(III) may be found, especially due to biologically mediated redox reactions (ratios 0.1- 250; Cullen and Reimer 1989). Several reports confirm the predominance of As(III) in early spring with the beginning of algal bloom, before significant methylation (section 2.5.2.1) in late summer and autumn takes place and leads to a predominance of methylated As (Anderson and Bruland 1991; Cabon and Cabon 2000; Howard et al. 1995; Sanders 1985; Sohrin et al. 1997). Korte and Fernando (1991) were among the first to point out that As(III) in groundwater is likely to be more predominant than previously assumed. Inadequate sampling techniques with oxidation of As(III) to As(V) during storage might be the reason for the reported clear predominance of As(V) in most older reports (McCleskey et al. 2004; section 2.7.1). The British Geological Survey De-partment of Public Health Engineering (2001) found that 50-60% of all As in the Holocene alluvial aquifers in Bangladesh is As(III). Arsenite predominates in sedimentary aquifers including black shales in Taiwan (Chen et al. 1994) and makes up 60-90% of the total As in Holocene alluvial and lacustrine aquifers in Inner Mongolia (Smedley et al. 2001). For aquifers in Holocene basin fill sediments in the San Joaquin Valley, California As(III) was found to increase with increasing well depth (Fujii and Swain 1995).

In the presence of sulfides under reducing conditions the formation of arsenic sulfides (orpiment As2S3, realgar AsS, arsenopyrite) may limit the aqueous As(III) concentrations (Cherry et al. 1979; Savage et al. 2000; Figure 5).

Compared to As(V), As(III) was found to be about 18-50 times more toxic by Petrick et al. (2000) and 25-60 times more toxic by Morrison et al. (1989). Investigations on macrophages, one of the principal and most sensitive immune effector cells, showed a 100-fold lower toxicity for As(V) than for As(III) (Sakurai et al. 1998). Arsenite reacts with sulfurhydryl groups of amino acids such as cysteine and thereby inactivates a wide range of enzymes in intermediate metabolism (Squibb and Fowler 1983).

pK1 pK2 pK3 reference 9.23 12.13 13.4 Naumov et al. (1974)* 9.23 12.13 12.71 Sadiq et al. (1983) 9.22 11.01 13.4 Brookins (1988)

14.2 15.1 Nordstrom and Archer (2003)

pK1 pK2

pK3

HAsO32-

H3AsO30 AsO3

3-

H2AsO3-

22 Arsenic

Oxidation of arsenite to arsenate is generally considered a protective mechanism against enzyme deactivation. Landner (1989), however, reports that under low phosphorous conditions arsenate is at least ten times more toxic than arsenite for algae cultures. Arsenate replaces phosphate, indistin-guishable for biota, disturbing central regulatory processes in the biological system concerning genetic (DNA), hormonal (cAMP), energetic (ATP) or enzymatic (substrates and protein phos-phorylation) codes (Squibb and Fowler 1983). Only under conditions of high phosphorous concen-trations, the toxicity of arsenate is insignificant compared to arsenite. Cullen and Reimer (1989) proposed that with high phosphate concentrations cells switch to a more selective transport system for phosphate, reducing arsenate accumulation. Probably this is not really a switch in transport systems but rather a competition resulting in the uptake of the predominant compound. Considering switches in the transport systems, low phosphate concentrations are more likely to trigger phos-phate specific uptake. While at high phosphate concentrations the non-specific and fast phosphate inorganic transport system (Pit) is active, the energy-requiring, specific phosphate inducible system (Pst) works during phosphate depletion. It is about 100 times more specific for phosphate than for arsenate (Petänen et al. 2001).

Complexation with other ions is rarely observed for inorganic As. At fluoride concentrations > 1.5 mg/L, HAsIIIO3F- and AsIIIO3F2- might form (Crecelius et al. 1986). Lowenthal et al. (1977) report that in seawater Ca- and Mg-As(V) complexes can predominate but Nordstrom (2000) emphasizes the uncertainty of information about the significance of aqueous metal As(V) complexes because of a lack of measured thermodynamic equilibrium constants. Assuming divalent metal P(V) com-plexes as analogues association constants for divalent metal As(V) complexes were estimated. Us-ing these for a water sample from Bangladesh showed that free As(V) concentrations decreased by about 50% when divalent As(V) complexes were considered (Nordstrom 2000). Under reducing conditions, at neutral to alkaline pH, soluble thioarsenites, commonly referred to as AsIIIS3

3-, be-come important (Cullen and Reimer 1989). The hypothesis by Kim et al. (2000) of stable As car-bonate complexes (AsIII(CO3)2

-, AsIII(CO3)(OH)2-, AsIIICO3

+) formed during a leaching experiment with As sulfides in NaHCO3 solution, was disproved by Wilkin et al. (2003). They found a pre-dominance of mononuclear thioarsenites at sulfide concentrations > 10-4.3M at neutral pH.

2.5.2 Dissolved organic arsenic species

Both inorganic trivalent and pentavalent As compounds can be methylated. The process is not purely chemical but requires the involvement of a living organism and, presumably, the interven-tion of As within the metabolic pathways of the cells. Cullen and Reimer (1989) as well as Bentley and Chasteen (2002b) provide excellent reviews about the As biomethylation pathway. Abiotic As methylation is possible, e.g., in the presence of methyliodide, but is rarely observed.

Early studies on organisms that methylated As focused on in vitro experiments using crude extracts of fungi and bacteria. Among the most prominent fungi for methylation of inorganic arsenicals are Apiotrichum humicola, formerly named Candida humicola (Cox and Alexander 1973; Cullen et al.

Species 23

1995), and Scopulariopsis brevicaulis (Challenger et al. 1933, 1954). The aerobic fungus Scopu-lariopsis brevicaulis also metabolizes Se and Te (Challenger 1951), as well as Sb (Jenkins et al. 1998). Andrewes et al. (2000) found that in the presence of Sb(III), As methylation from As(III) by Scopulariopsis brevicaulis is inhibited significantly, even though in the absence of Sb, As is much more readily biomethylated than Sb in the absence of As (1.2-5.3% of the added As biomethylated compared to 0.0006-0.008% of the added Sb). On the other hand, (CH3)3Sb production increased with increasing dissolved As(III) concentrations. Several fungi also have a more selective methyla-tion mechanism. Gliocladium roseum and Penicillium notatum e.g., do not methylate inorganic As, but readily metabolize alkylarsenicals (Cox and Alexander 1973).

Although Challenger (1945) doubted that bacteria can methylate As, McBride and Wolfe (1971) were the first to prove the production of (probably) (CH3)2AsH by a Methanobacterium strain MoH (Archaeobacter group). Wickenheiser et al. (1998) also confirmed the production of volatile inor-ganic and organic As from Methanobacterium formicicum. A number of nonmethanogenic bacteria have also been identified as methylarsine producers, e.g., Pseudomonas and Flavobacterium (Shariatpanahi et al. 1981, 1983), Escheria coli (Shariatpanahi et al. 1981), Proteus, Achromobac-ter, Aeromonas, Enterobacter (Shariatpanahi et al. 1983), as well as several bacteria not normally associated with arsenic metabolism, e.g., Veillonella alcalescens, Streptococcus sanguis, and Fuso-bacterium nucleatum (Pickett et al. 1988). A detailed review of the microbiological methylation of As is provided in Cullen and Reimer (1989).

Temperature seems to be one decisive factor in the methylation process. Howard et al. (1982, 1984) described a temperature threshold, stating that no methylation occurs below 9-12°C. The methyla-tion reaction pathway as indicated already in Figure 3 is a sequence of alternating reduction steps (including the replacement of OH--groups by CH3

--groups) and oxidizing steps. The principle was already proposed by Challenger (1945), however, detailed information about the compounds and enzymes involved ( Figure 7) was only gathered in more recent research studies on laboratory ani-mals or humans.

As described above (section 2.5.1.2), inorganic As(V) is first reduced either non-enzymatically by gluthathione (GSH) or other thiols (RSH) or enzymatically by ArsC, arsenate reductase (Aposhian 1997; Wildfang et al. 1998), to inorganic As(III), then methylated to Monomethylarsonic acid (MMAVA, (CH3)AsVO(OH)2). Concerning the nature of the methyl donor, Challenger (1945) pro-posed another already methylated compound such as a betaine, a methionine, or a choline deriva-tive. Experiments with 14CH3 labeled derivatives revealed that only methionine transferred its methyl group to As(III) and was detected in the final volatile product (CH3)3As. Cantoni (1953) identified the methionine derivative as S-adenosylmethionine (SAM). Yet, it was another 40 years later that Cullen et al. (1995) actually proved the direct uptake of deuterium labeled methionine also by non-volatile As metabolites in growing cultures of Apiotrichum humicola.

Thus, MMAVA is metabolized by the conversion from SAM to SAHC (S-adenosylhomocysteine) in the presence of the enzyme arsenite methyltransferase (Wildfang et al. 1998; Zakharyan et al.

24 Arsenic

1999). For further methylation, MMAVA has to be reduced to Monomethylarsonous acid (MMAIIIA, (CH3)AsIII(OH)2) by RSH and the enzyme MMAV reductase (Zakharyan and Aposhian 1999). MMAIIIA is then methylated to dimethylarsinic acid (DMAVA, (CH3)2AsVO(OH)) via the conversion from SAM to SAHC and in the presence of MMAIII methyltransferase (Zakharyan et al. 1999).

AsVO(OH)3 AsIII(OH)3 Arsenate Arsenite

(CH3)AsVO(OH)2 (CH3)AsIII(OH)2 Monomethylarsonic acid Monomethylarsonous acid

(CH3)2AsVO(OH) (CH3)2AsIII(OH) Dimethylarsinic acid Dimethylarsonous acid

(CH3)3AsVO (CH3)3AsIII ↑ Trimethylarsineoxide Trimethylarsine (b.p. 52°C)

Figure 7 Model for the As methylation pathway; GSH = glutathione, RSH = other thiols, RS-SR = disulfide that is recycled to RSH by RSH reductase, SAM = S-adenosylmethionine, SAHC = S-adenosylhomocysteine, b.p. = boiling point, (GS)2AsSe- = seleno-bis(S-glutathionyl) arsinium ion (references: general scheme already proposed by Cullen and Reimer 1989, enzymes purified by 1 Aposhian 1997; Wildfang et al. 1998; 2 Wildfang et al. 1998; Zakharyan et al. 1999; 3 Zakharyan and Aposhian 1999; 4 Zakharyan et al. 1999)

Compared to arsenite methyltransferase (Michaelis constant Km = 5.5⋅10-6 M in rabbit liver) and MMAIII methyltransferase (9.2⋅10-6 M), the enzyme MMAV reductase (2.16⋅10-3 M) shows the weakest enzyme-substrate affinity, and is thus the rate-limiting enzyme (Zakharyan and Aposhian 1999). This rate-limiting process can explain why not only the end products, but also intermediates in the As biomethylation pathway can be found in different studies, e.g., MMAIIIA (section 2.5.2.3). According to Zakharyan and Aposhian (1999) it is also possible to reduce MMAV in the absence of the enzyme only with GSH, however this non-enzymatic chemical reduction yielded 11 times lower reaction rates. On the other hand, the MMAV reductase enzyme absolutely required the presence of GSH. Neither L-cysteine, nor DTT (dithiothreitol) proved to be sufficient as alternative reductants. L-cysteine, however, yielded greater activities for As(III) and MMAIII methyltrans-ferase in in vitro experiments with hamster liver (Wildfang et al. 1998) and dithiols were found to be more effective for MMAIII methyltransferase from hamster liver (Zakharyan et al. 1999).

Arsenate Reductase1

Arsenite Methyltransferase2

MMAIII Methyltransferase4

MMAV Reductase3

SAM

SAM

SAHC

SAHC (GS)2AsSe-

GSH RSH

RS-SR

RSH

RS-SR

RSH

RS-SR

RSH

RS-SR

Species 25

Aposhian et al. (1999) report from in vitro experiments that methylation by methylcobalamin (me-thylvitamin B12) only requires a reducing environment but no enzymatic reactions. This report might indicate a different metabolic pathway for anaerobic microorganisms, a controversial subject as presented in detail in Bentley and Chasteen (2002b). Rosen (2002) reviews differences in spe-cific proteins resulting from separate evolutionary pathways between prokaryotes and eukaryotes. The overall scheme of As biomethylation, however, seems to be valid as explained above.

The majority of research groups postulates DMAVA to be the ultimate metabolite in humans (Buchet et al. 1981; Hopenhayn-Rich et al. 1996b; Le et al. 2000a), while it is further methylated to

trimethylarsineoxide (TMAVO, (CH3)3AsVO) by the fungus Apiotrichum humicola (Cullen et al. 1995), and hamsters and rats (Yamauchi and Yamamura 1984a; Yoshida et al. 1997). Styblo et al. (1999b) did not detect metabolism of DMA compounds for rat hepatocytes. The TMAVO concen-trations of several ng/L in all urine samples investigated by Sur (1999) questioned DMAVA as the ultimate metabolite and supported the hypothesis of further methylation beyond DMAVA. Marafante et al. (1987) made a similar observation, 48 hours after application of 0.1 mg DMAA about 4% of the applied dose was excreted as TMAVO.

Trimethylarsine (TMA, (CH3)3AsIII) is the theoretical volatile end product in the outlined reaction pathway. During investigations of intra-oral air from 6 test persons, however, neither TMA, nor mono- (MMA), nor dimethylarsine (DMA) were detected in human breath as volatile intermediates in the As biomethylation pathway (Feldmann et al. 1996).

One factor that might inhibit the methylation process is the formation of the seleno-bis(S-glutathionyl) arsinium ion (GS)2AsSe- ( Figure 7). Aposhian et al. (1999) and Gailer et al. (2000) propose that (GS)2AsSe- is rapidly formed from arsenite in the presence of GSH, before any methy-lation occurs, and is easily excreted. That can explain the known antagonism between As and Se which was first reported in the late 1930s (Moxon 1938). When co-administered Se and As mutu-ally inhibit the individual methylation pathways. Lethal doses for co-administration of Se and As are higher than for administration of either of the elements alone.

Aposhian et al. (1999) report a lower limit for methylation that could also explain the known delay in excretion of methylated As compounds after As exposure (Vather 1999). According to studies on rabbit livers, arsenite firmly binds to proteins (20 times better than arsenate). It seems that me-thylation does not take place before the protein binding sites are occupied and additional arsenite accumulates. Concerning an upper limit for methylation, there seems to be some evidence that me-thylation decreases with very high concentrations of inorganic As, indicating an inhibition of the methylation process by an excess of As(III) and a saturation of the enzymic conversion from MMAA to DMAA (threshold hypothesis by Petito and Beck 1990). Hopenhayn-Rich et al. (1993) conclude from epidemiological and experimental human studies cited in the literature that those data do not support the methylation threshold hypothesis, because there was always about 20-25% non-methylated inorganic As present in the urine regardless of the absorbed dose of inorganic As. In contrast, results from Farago and Kavanagh (1999) show 79-96% DMAA in the urine of popula-

26 Arsenic

tions at low exposure levels compared to 60% at high exposure levels, supporting the threshold hypothesis. Also an increase of MMAA compared to DMAA can be taken as evidence for a distur-bance or saturation of the methylation process, as shown by Styblo et al. (1999a, 1999b) for human and rat hepatocytes. They also found a decrease of total methylation at higher As concentrations, suggesting the inhibition of methylation according to the threshold hypothesis. An example for methylation inhibition in environmental samples might be Mono Lake that has inorganic As con-centrations of up to 17 mg/L, but no methylated As species (Anderson and Bruland 1991).

Like all organometallics, organic As compounds are thermodynamically unstable. However, their decomposition kinetics is so slow that they may have a transient existence under a variety of envi-ronmental conditions. For microbial demethylation from DMAA to As(V) by an estuarine micro-bial culture, Sanders (1979) reports a rate of 1 ng/L/day. Chemical demethylation is very slow. However, oxidation, reduction, hydrolysis, and reactions with sulfur compounds can significantly alter distribution, volatility, and mobility of biologically produced compounds.

2.5.2.1 Mono- and dimethylated As(V) acids

As explained above, MMAVA [(CH3)AsVO(OH)2] and predominantly DMAVA [(CH3)2AsVO(OH)] are the main metabolites in the As biomethylation pathway, except for those animals who lack the As methyltransferase enzymes such as the guinea pig (Healy et al. 1997), gorilla, orangutan, and chimpanzee (Wildfang et al. 1997). Several studies indicate an average of 10-30% inorganic As, 10-20% MMAA and 60-70% DMAA in human urine after exposure to inorganic As (Vather 1999). Compared to inorganic As, MMAVA and DMAVA are much more easily excreted because they are less reactive with tissue constituents. Following oral intake of methylated As compounds, about 75% of the ingested dose is excreted after 4 days, compared to only 45% of a similar dose of inor-ganic As (Buchet et al. 1981). Vather (1999) reviews variations in the human metabolism of As and shows differences between different population groups (the percentage of MMAA in urine was detected to be below average with only 2-4 % in Andean natives, for people from Taiwan above average with 27%) and individuals, also depending on age (children seem to have a lower methyla-tion capacity and a higher As retention potential; Del Razo et al. 1999), gender (women showed 3% more DMAA, but less MMAA in urine compared to men, pregnant women up to 90% DMAA), exposure level (see threshold hypothesis discussed above), and nutritional status (lack of vitamins and methyl donors may inhibit As metabolism). However, detailed studies are still rare and it is often hard to exclude cross reactions and factors influencing each other.

According to a review on As species in plants (Meharg and Hartley-Whitaker 2002), rice, apple, tomato (in a laboratory experiment with a deficiency of nitrogen and phosphorous), wild straw-berry, bilberry, raspberry (only DMAA), Norway spruce, white alder, sedge, and cedar were found to contain significant amounts of MMAA and DMAA (no distinction between penta- and trivalent species). However, it is not yet clear, whether plants actually methylate As themselves or the As is methylated by microbes in the rhizosphere and then taken up by the plant.

Species 27

Extracellular MMAA and DMAA (without distinction between penta- and trivalent species) were already determined together with arsenate and arsenite in 1973 in some fresh and seawater samples by Braman and Foreback (1973). While the average amount of methylated compounds (predomi-nantly DMAA), is 10-20% of the total As, it could also become up to 70% depending on bioactiv-ity and temperature. These results compare well with several later studies, e.g., Sanders (1985), who determined up to 60% methylated compounds during the summer bloom in an estuary (Chesa-peake Bay). The concentrations of MMAA in this study were equal to or even exceeded those of DMAA. For the estuarine Southampton Waters, United Kingdom, where phytoplankton concentra-tions are about 10 times less than those in Chesapeake Bay, DMAA made up about 25% of total As, MMAA about 12% (Howard et al. 1995). Andreae (1978) and Froehlich et al. (1985) deter-mined the proportion of methylated compounds in freshwater samples to average 10% of total As, with MMAA concentrations being only 1/10 of those of DMAA. Anderson and Bruland (1991) reported between 1 and 59% methylated As in Californian lakes during late summer and fall with a clear predominance of DMAA. Sohrin et al. (1997) detected up to 64% DMAVA in Lake Biwa, Japan, in summer, MMAVA was an order of magnitude less. Cabon and Cabon (2000) found 20% DMAA and 3% MMAA during the phytoplankton bloom at the French Atlantic. Methylated As concentrations of less than 10% total As (DMAA > MMAA) determined in subarctic lakes in Can-ada are below the average values described above and seem to support the temperature dependence of methylation mentioned in section 2.5.2. Interestingly, all these methylarsenicals were found in oxic waters, in contrast to the publication of Wood (1974), and the still generally widespread opin-ion that methylarsenicals would only be found under strictly anoxic conditions. The possibility that such reducing conditions occur in microenvironments is certainly not discounted.

Lower concentrations of methylated compounds (0.01-1% of the total As) were found in geother-mal waters (Hirner et al. 1998). From 6 samples, 5 showed a predominance of MMAA. Only in one sample (the one with the lowest temperature, 20°C) DMAA concentrations exceeded MMAA by about one order of magnitude. The compounds MMAA and DMAA were also detected in German harbor and river sediments by Krupp et al. (1996) and in deposited domestic waste and contami-nated soil in the vicinity of a gas station and a coal processing industry (Hirner et al. 2000). Bednar et al. (2004) reported concentrations of up to 100 µg As/L for MMAA, and 10 µg As/L for DMAA in the stagnant surface water of a cotton-producing area in Arkansas that might have been impacted by the use of monosodium methylarsonate as cotton herbicide. No methylated compounds were found in the groundwater.

Figure 8 shows the predominant MMAVA species as a function of pH. Similar to inorganic As(V), the negatively charged complex CH3AsO2(OH)- predominates under near neutral conditions, and the uncharged complex only in acid solutions (pH < 3.6).

Figure 9 shows the predominant DMAVA species as a function of pH. According to Doak and Freedman (1970) as well as Baes and Mesmer (1976), only two species exist for DMAVA, and the predominant one (pH < 6.2) is the uncharged (CH3)2AsO(OH).

28 Arsenic

(CH3)AsO(OH)2

0 ↔ (CH3)AsO2(OH)- + H+ ↔ (CH3)AsO3

2- + 2H+

Figure 8 pH-dependent dissociation of MMAA species (a(As) = 1.33⋅10-7 mol/L; * data used for the diagram)

In contrast to the highly mobile inorganic As(III) uncharged complex, the uncharged DMAVA complex shows some non-specific sorption both on cation and anion exchangers. This non-specific sorption could be caused by hydrophobic interactions with the exchanger matrix that the hydro-philic inorganic As(III) does not exhibit. Korte and Fernando (1991) and Hansen et al. (1992) sug-gested a protonated DMAVA species (CH3)2AsO(OH)-H+ that appears below pH 4 and becomes predominant at pH < 1.6, and pH 1.2 respectively.

(CH3)2AsO(OH)0 ↔ (CH3)2AsOO- + H+

Figure 9 pH-dependent dissociation of DMAA species (a(As) = 1.33⋅10-7 mol/L; * data used for the diagram)

MMAVA and DMAVA are the only dissolved methylated As species positively identified extracel-lular in environmental samples so far.

2.5.2.2 Tri- and tetramethylated As(V) acids

Less frequent than MMAVA and DMAVA is TMAVO [(CH3)3AsVO]. The pK value for the reaction (CH3)3AsVO ↔ (CH3)3As+OH + H+ is 3.6 (Hansen et al. 1992). The rare occurrence of TMAVO might be due to facile reduction to the volatile (CH3)3As by aerobic and anaerobic organisms, as well as in the presence of thiols (RSH). TMAVO has so far mainly been found intracellular in a number of fish species. While Cullen and Dodd (1989) describe TMAVO as a minor component, maybe even the breakdown product of some yet unknown precursor only, Kaise et al. (1997) found it to be predominant in fish with 55-78% of total As. Also marsh snails contained significant

pK1 pK2 reference 3.61 8.24 Doak and Freedman (1970)* 3.6 8.2 Baes and Mesmer (1976) 4.1 8.7 Korte and Fernando (1991)

pK1 pK2 reference --- 6.27 Doak and Freedman (1970)* --- 6.2 Baes and Mesmer (1976) 1.2 --- Hansen et al. (1992) 1.6 6.3 Korte and Fernando (1991)

pK1 pK2

pK2

(CH3)2AsO(OH)0 (CH3)2AsOO-

(CH3)AsO(OH)20

(CH3)AsO2(OH)-

(CH3)AsO32-

Species 29

amounts of TMAVO. In their review on As species in plants, Meharg and Hartley-Whitaker (2002) list cocksfoot grass, Norway spruce, European Larch, wild strawberry, and raspberry as plants con-taining TMAVO.

Kaise et al. (1997) also claim to have found 2 µg/L TMAO (6.8% of total As) extracellular in the Hayakawa river at hot springs in Hakone, Japan. Hirner et al. (1998) report 0.18 and 0.25 µg/L TMAO for two samples of geothermal waters from New Zealand. However, both studies used hy-dride generation under the oversimplified assumption that only TMAO can produce the volatile TMA (section 2.6.1.1). Similarly questionable is the detection of TMAO as TMA in pore water of subarctic lake sediments in Canada (Bright et al. 1996), in German harbor and river sediments by Krupp et al. (1996), and in deposited domestic waste and contaminated soil in the vicinity of a gas station and a coal processing industry (Hirner et al. 2000). No evidence is found so far for the exis-tence of TMAO in normal freshwater environments.

The tetramethyl arsonium ion (CH3)4As+ (TETRA) has only been found in tissues of clams (Mere-trix lusori; Shiomi et al. 1987), a crab, a sea cucumber (Shibata et al. 1992), and other gastropods (Francesconi et al. 1988), never extracellular in natural waters. Koch et al. (1999) report trace amounts of TETRA in the monkey flower Mimulus sp., thought to be synthesized by this plant. Wild strawberry, broad buckler fern, and cocksfott grass are other examples of plants with TETRA occurrences (Meharg and Hartley-Whitaker 2002). Geiszinger et al. (2002) report for the first time a predominance of TETRA (up to 85% of the arsenate accumulated from seawater) in marine anne-lides (Polychaetes Nereis diversicolor and Nereis virens). The biosynthesis of TETRA is unclear because the end product of As biomethylation according to Figure 3 is (CH3)3As. Tetramethyl ar-sonium could result from decarboxylation of arsenobetaine (section 2.5.2.4), however Geiszinger et al. (2002) found no evidence of such a transformation in their study.

2.5.2.3 Mono- and dimethylated As(III) acids

The existence of the trivalent organic acids MMAIIIA [(CH3)As(OH)2] and DMAIIIA [(CH3)2As(OH)] was neglected or doubted in many former reviews (e.g., Farago and Kavanagh 1999; Korte and Fernando 1991; Vather 1999). Cullen and Reimer (1989) mentioned them as pos-sible theoretical intermediates in the biotransformation pathway, but also stated that they were "un-known". Aposhian et al. (2000), Le et al. (2000a, 2000b), Sampayo-Reyes et al. (2000), Wildfang et al. (1998), and Zakharyan and Aposhian (1999) proved their existence as metabolites in mam-mals exposed to increased concentrations of inorganic As. In all these reports MMAIIIA clearly predominated over DMAIIIA. Aposhian et al. (2000) determined the share of MMAIIIA in urine of humans exposed to inorganic As in drinking-water to be 7-11% of the total As and 35-44% of the total MMAA (MMAIIIA + MMAVA). It seems that a certain amount (in the millimolar range) of MMAVA reductase, the rate-limiting enzyme for the biotransformation, is needed for a significant conversion from MMAVA to MMAIIIA. More recent reports also emphasize the probability of erro-neous interpretation of all MMAA just as MMAVA in earlier studies without actually measuring MMAIIIA and the consequent need of re-interpretation of some of these data.

30 Arsenic

Extracellular occurrence of trivalent methylated compounds is reported for Lake Biwa, Japan, by Hasegawa et al. (1994) with 0.3-3.0 nM MMAIIIA and 0.2-11 nM DMAIIIA, and by Sohrin et al. (1997) with < 20 ng/L MMAIIIA and DMAIIIA. Gong et al. (2001) showed that the few reports on detection of trivalent methylated species, especially DMAIIIA, are probably caused by inadequate sampling techniques. When stored at room temperature, MMAIIIA was completely oxidized within one week and DMAIIIA in 17 hours. With storage at -20°C only 10% MMAIIIA was oxidized over a period of 5 months, while DMAIIIA was found to be very unstable showing complete oxidation within 15 days. Stability in urine is even less, with 98% MMAIIIA oxidized after 30 days and 100 % DMAIIIA after 90 min at room temperature.

Several papers state that As methylation in contrast to Hg methylation is always a detoxification mechanism, with methylated As compounds being several 100 times less toxic than inorganic As compounds (e.g., Korte and Fernando 1991; Morrison et al. 1989; Vather 1999; Yamauchi and Fowler 1994). The pentavalent organic As compounds might in fact be less toxic than the inorganic arsenicals. Sakurai et al. (1998) report for macrophages a 1000-fold lower toxicity for DMAVA compared to As(III) and no toxicity at all was detected for MMAVA or TMAVO. Whereas Mass et al. (2001) could not determine any DNA nicking or degrading up to MMAVA concentrations of 3 M and DMAVA concentrations of 300 mM, Yamanaka et al. (1997) reported for laboratory ani-mals that DMAVA might be able to damage DNA and promote tumors. In their overview about genotoxicity of organoarsenic compounds Florea et al. (2004) show that DMAVA so far has proven to be responsible for cell growth inhibition, decrease of cell viability, apoptosis (cell death), sister chromatide exchanges, carcinogenity, and mutagenicity.

Even though the pentavalent organic arsenicals might be less toxic under certain aspects, already Horiguchi (1970) claimed that the intermediates in the As biomethylation pathway, MMAIIIA and DMAIIIA, are more toxic to mammals than inorganic arsenite. Petrick et al. (2000) confirmed for human hepatocytes that MMAIIIA is up to 26 times more toxic than arsenite and in fact the most toxic compound in the biotransformation pathway. Investigations of rat and human cells by Styblo et al. (2000) revealed a 4-20 times higher cytotoxicity for MMAIIIA compared to As(III), whereas the cytotoxicity of DMAIIIA was comparable to that of As(III). Comparable to that, for rat hepato-cytes MMAIIIA was about 25 times more toxic than As(III) or DMAIIIA (Styblo et al. 1997). Inter-esting is the study of Sakurai et al. (1998) who report an increase of macrophage cell toxicity for DMAA when GSH is added. Because GSH is the major reductant in the As biomethylation path-way ( Figure 7) this result supports the conclusion that DMAVA was further reduced to the more toxic DMAIIIA. Although this hypothesis was also discussed by the authors it was not investigated further.

The higher toxicity of the trivalent organic As species is related to their high affinity for specific cellular proteins and their potential as inhibitors for enzymes which generate cellular energy in the citric acid cycle (Styblo et al. 1997). The MMAIIIA has an additional toxic potential in that it is retained in cells and, in contrast to DMAIIIA, not released into the medium. According to Lin et al. (1999), MMAIIIA is over 100 times more potent than inorganic As(III) as an inhibitor for thiore-

Species 31

doxin reductase, a selenoprotein for the reduction of disulfides and therefore critical for the main-tenance of the intracellular redox state. Disturbances to the reduction of thioredoxin, a modulator of cell proliferation and DNA synthesis, might result in tumor growth. Mass et al. (2001) found that MMAIIIA and DMAIIIA were able to nick DNA in vitro without the need of exogenously added enzymatic or chemical activation at concentrations of 30 mM, and 150 µM respectively. Full DNA degradation occurred at 60 mM MMAIIIA, and 10 mM DMAIIIA respectively. In single-cell gel assays in human lymphocytes, MMAIIIA was 77 times, DMAIIIA 386 times more potent than inor-ganic As(III) in damaging the DNA.

Little is known about the toxicity of TETRA. According to Penrose (1974) it is harmless. Maeda (1994) specified that it is excreted renally without undergoing any metabolism. In contrast to this, Shiomi (1994) states that it is considerably lethal to marine organisms, being more toxic than MMAA and DMAA.

2.5.2.4 Arsenosugars, arsenocholine, arsenobetaine

Dimethylarsinoyl ribosides, better known under their trivial name arsenosugars X - XIII ( Figure 10), were found as intermediates of the As cycling in marine ecosystems, mainly in algae (Ed-monds and Francesconi 1981; Shibata et al. 1992) and seaweeds (Edmonds et al. 1982; McSheehy et al. 2000). Arsenosugar XIII was the major metabolite of the marine diatom Chaetoceros con-cavicornis (Edmonds et al. 1997). Shibata et al. (1992) also found arsenosugar XI and, more sel-dom, arsenosugar X in bivalves, and gastropodes. Samples from freshwater or terrestrial environ-ments are rare. Lai et al. (1997) found the terrestrial soil / freshwater cyanobacterium Nosto sp. to contain up to 32% arsenosugar X. Koch et al. (1999) reported for the Meager Creek hot springs in British Columbia the occurrence of the arsenosugars X and XI in microbial mats (maybe including cyanobacteria) and green algae as well as of arsenosugar X in moss and arsenosugar XI in lichens.

With: X R1 = OH, R2 = OH

XI R1 = OH, R2 = O-P-O-CH2

XII R1 = OH, R2 = SO3H R1 = NH2, R2 = SO3H

XIII R1 = OH, R2 = OSO3H

Figure 10 Structure for arseno-sugars and abbreviations for the different structures (numbering arsenosugar X to XIII according to Shibata et al. 1992)

O

OH CHOH

CH2OH

OH OH

H H H H

C C

C C

O

OCH2CHCH2-R2 O (CH3)2As-CH2

R1

32 Arsenic

Arsenosugars are stable at neutral pH and react with acids or bases to DMAVA. Consequently, DMAVA increases in human urine following arsenosugar ingestion (Le and Ma 1998). Arseno-sugars, especially arsenosugar XI, are probably the link between hydrophilic and lipophilic As compounds. Those sugarderivatives are anaerobically degraded to dimethyloxarsylethanol (CH3)2AsOCH2CH2OH that is the intermediate for arsenocholine (AC, (CH3)3As+CH2CH2OH) ( Figure 11). The existence of AC was claimed in shrimps by Norin et al. (1983) and in one gastro-pod by Shiomi et al. (1987).

Figure 11 Interactions of the organic As species arsenosugar, arsenocholine, and arsenobetaine

The (CH3)2AsOCH2CH2OH is easily converted by further oxidation to (CH3)2AsOCH2COOH, the intermediate for arsenobetaine (AB, (CH3)3As+CH2COO-) ( Figure 11). The pK value for the reac-tion (CH3)3As+CH2CO(OH) 0 ↔ (CH3)3As+CH2COO- + H+ is 2.18 (Hansen et al. 1992). AB was first isolated in 1977 (Edmonds et al. 1977). It is the predominant As species in marine organisms that generally show organoarsenical concentrations significantly above average background (≈ 2 ppb; Penrose 1974; Shibata et al. 1992). This predominance is commonly related to accumulation of compounds that have been synthesized from arsenate at lower trophic levels (Ünlü and Fowler 1979), even though investigations from Klumpp and Peterson (1981) show that organoarsenicals in snails do not come from the lower trophic level (algae), but are manufactured directly from arse-nate. AB is very stable and excreted from human bodies without metabolism (Crecelius 1977b; Ma and Le 1998). Only under extreme temperatures (150-180°C) it decomposes into TMAVO and TMA (Devesa et al. 2001). Neither MMAA, DMAA, TMAO, TETRA, nor AC showed these de-compositions with temperature. AB is essentially nontoxic (Cullen and Reimer 1989; Francesconi and Edmonds 1997).

2.5.2.5 Organic arsenic sulfur compounds

The As biomethylation pathway is further complicated by the ability of As compounds to react with thiols (Bentley and Chasteen 2002b; Delnomdedieu et al. 1994). Complexation of inorganic As(III) with GSH produces arsenotriglutathione AsIII(GS)3 which is easily excreted at the beginning of the As biomethylation pathway, comparable to (GS)2AsSe- (section 2.5.2). It is interesting to remark that AsIII(GS)3 inhibits the GSH reductase enzyme that recycles GS-SG to GSH. It could thus be a potent inhibitor for any further As methylation. For organic As-S-compounds the follow-ing sequence is likely (Bentley and Chasteen 2002b; Cullen et al. 1984):

arsenosugars (CH3)2AsOCH2CH2OH (CH3)2AsOCH2COOH

(CH3)3As+CH2CH2OH (CH3)3As+CH2COO-

arsenocholine arsenobetaine

oxidation anaerobic decomposition

oxidation

reduction + CH3

reduction + CH3

Species 33

(CH3)xAsVO(OH)3-x + 2RSH → (CH3)xAsV(SR)2(OH)3-x + H2O (CH3)xAsV(SR)2(OH)3-x → (CH3)xAsIII(OH)3-x + RSSR (CH3)xAsIII(OH)3-x + (3-x)RSH → (CH3)xAsIII(SR)3-x + 2H2O for x = 1 (monomethyl), and x = 2 (dimethyl compounds)

Reduction of trimethylarsine oxide to trimethylarsine was possible with a variety of thiol reagents (cysteine, dimercaptopropanol, lipoic acid, mercaptoacetic acid, and mercaptoethanol). The follow-ing reactions were postulated by Bentley and Chasteen 2002b:

(CH3)3AsVO + RSH → (CH3)3AsV(SR)(OH) (CH3)3AsV(SR)(OH) + H3O+ → (CH3)3AsIII(SR)+ + 2 H2O (CH3)3AsIII(SR)+ + RSH → (CH3)3AsIII(SR)2 + H+ (CH3)3AsIII(SR)2 → (CH3)3AsIII ↑ + RS-SR

In anoxic environments, with pH < 7, and in the presence of H2S or thiols (RSH) methylated As-S-compounds could also form naturally. During periods of high bioproductivity, thiol concentrations can reach values comparable to those of inorganic S-compounds and the presence of such As-thiol-compounds would lead to an underestimation of the extent of biomethylation (Cullen and Reimer 1989). Like the (CH3AsIIIO)x compounds, the (CH3AsIIIS)x compounds are more toxic than inor-ganic As(III). CH3As(SR)2, is two orders of magnitude more potent as an inhibitor of glutathione reductase, the key enzyme in the detoxification pathway to DMAVA, than As(III) (Styblo et al. 2000). Between (CH3AsIIIO)x and (CH3AsIIIS)x, Cullen et al. (1989) state that for the fungus Apiotrichum humicola (CH3AsIIIS)x is more toxic, whereas Styblo et al. (2000) found that (CH3)2As(SR) is less toxic to human cells than (CH3)2As(OH). The existence of thioorganoarsenate as a mammalian metabolite was recently proven by Hansen et al. (2004). A 2-dimethylarsinothiol acetic acid was characterized in the urine of Orkney Island wild sheep that consume about 30 mg As as arsenosugars daily from their major food source, seaweed.

Bright et al. (1996) deduct the existence of (CH3)xAsIII(SR)3-x from the formation of volatile methyl compounds at pH 6 during hydride generation on pore water samples from subarctic lake sediments in Canada. Whereas MMAVA, DMAVA, and TMAVO are not volatilized above pH 1 (section 2.6.1.1), monomethyl-, dimethyl-, and trimethyarsinothiols were proven to do so. Because trimethylated arsenicals were not considered in this study, it is not really clear whether the original substances could also have been MMAIIIA and DMAIIIA that volatilize at a higher pH (pH > 5). This study provides the only “evidence” for the existence of (CH3)xAsIII(SR)3-x in aqueous samples.

2.5.2.6 Organic arsenic halogen compounds

De Bettencourt et al. (1994) appear to be the first to propose the possible existence of dissolved dimethyl halogenated As species, (CH3)2OAsCl (or (CH3)3OHAsF), in natural water samples from the Tagus estuary in Portugal. They made up 19-25 % of the total As species. The results were deduced from mass spectra after DCI (desorption chemical ionization)-MS/MS analysis ( Figure

34 Arsenic

12). However, the compound detected proved refractory to normal hydride generation techniques by not forming volatile As species, a fact that is hardly explicable from the chemical structure of a dimethyl chloro or a trimethyl fluoro As compound. Alternatively, the authors interpreted the de-tected compound as halo-arsenobetaine, or halo-arsenocholine.

Figure 12 DCI mass spectrum for an As species assumed to be (CH3)2OAsCl, (CH3)3OHAsF, or a halogenated arsenobetaine (de Bettencourt et al. 1994; % = rela-tive abundance by taking pre-dominant mass = 100)

Hasegawa et al. (2002) investigated the conversion of inorganic and methylarsenic(III) hydroxides into the corresponding halides in aqueous HCl solutions and determined stability constants and partition coefficients ( Figure 13).

Figure 13 Dissociation in HCl solution accord-ing to Hasegawa et al. (2002) of (a) inorganic As(III) <1⋅10-3 mol/L, (b) MMAIIIA <1⋅10-3 mol/L, (c) DMAIIIA 1.0⋅10-4 mol/L, ((CH3)2As)2O is a dimethyl As dimer, the so-called cacodyloxide that forms at low total DMAIIIA concentrations; (d) DMAA 1.0⋅10-8 mol/L; AsCl3, (CH3)AsCl2, and (CH3)2AsCl are volatile species and will be dis-cussed in section 2.5.3.6

(CH3)2AsOCl ?Halogenated arsenobetaine?

0

10

2 0

3 0

4 0

5 0

6 0

7 0

8 0

9 0

10 0

5 0 7 0 9 0 110 13 0 15 0 17 0 19 0m / z

139

157

159 121

105

rela

tive

abun

danc

e [%

]

Species 35

The stepwise substitution of hydroxide ions by chloride ions can be described as (CH3)nAs(OH)i+1Clj-1 + H+ + Cl- ↔ (CH3)As(OH)iClj + H2O (n + i + j = 3). The stability constants increase from As(III) to MMAIIIA to DMAIIIA. Methyl groups are classified as weak electron do-nors. Attached to the As atom they increase the electron density on the central As atom and reduce the ionic interaction of As-OH bonds. Thus, a higher number of methyl groups facilitates the re-placement of hydroxide ions by chloride ions bound to the As, and organic As-Cl species occur at lower HCl concentrations (higher pH) than inorganic As-Cl species ( Figure 13).

2.5.3 Volatile arsenic species

Volatile arsenicals, so-called As hydrides, are As compounds with a boiling point below 150°C. They can be formed either as intermediates or end products of the As biotransformation pathway as indicated in Figure 3 or in a purely chemical reaction. This chemical hydride generation is com-monly used in analytical techniques; details will be discussed in section 2.6.1.1.

The first indication about the existence of volatile As is cited in Gmelin (1839). After several inci-dents of arsenic poisoning in Germany and Great Britain, Gmelin noted that a garlic smell was present in the rooms, in which arsenic poisoning occurred. Following this clue, Fleck (1872) showed experimentally that molds growing on the wallpapers that were painted with As pigments (copper aceto-arsenites, called Scheel's green, and Schweinfurt or emerald green, used from 1806 until the 1960s), were able to produce a volatile, toxic and garlic smelling arsenic compound. Fleck proposed the produced gaseous compound to be arsine (AsH3) but he was not able to identify it. Finally, Gosio was able to isolate various cultures that produced this gas and it was named Gosio gas in his honor (Gosio 1893, 1901). Biginelli (1900a, 1900b) purged the Gosio gas produced by the molds through acidified HgCl solution and identified the resulting precipitate, incorrectly, as diethylarsine (C2H5)2AsH. About 30 years later Challenger et al. (1933) reviewed Biginelli´s work and positively identified the produced gas as trimethylarsine. The mold, earlier named by Gosio Penicillium brevicaule was later identified as Scopulariopsis brevicaulis (Challenger 1945).

2.5.3.1 Inorganic AsH3

Although Scopulariopsis brevicaulis is one of the most prominent fungi in the methylation process of inorganic to organic (volatile) As, it does not produce AsH3. According to Cheng and Focht (1979), Pseudomonas and Alcaligenes (Achromobacter) produce AsH3 both from arsenite and ar-senate under anaerobic conditions. Michalke et al. (2000) incubated sewage sludge anaerobically and detected 0.76 ng/L AsH3. The highest concentrations of all volatile As compounds were found in pure cultures of the archaea Methanobacterium formicicum, whereas the Methanobacterium thermoautotrophicum produced AsH3 only. According to the studies of Michalke et al. (2000) sul-fate-reducing bacteria are not able to form AsH3.

With a boiling point of -55°C (Lido 2003), AsH3 is the most volatile of the As gases. It is unstable under atmospheric conditions and is readily oxidized by oxygen (50 µg are totally oxidized in air

36 Arsenic

within 120 hours; Pantsar-Kallio and Korpela 2000). The Henry constant for solubility in water at 298.15 K is 0.0089 mol/kg/bar (Wilhelm et al. 1977).

AsH3 is the most toxic inorganic form of As, causing immediate hemolysis. Fewer than 250 cases of arsine gas poisoning have been reported in the past 65 years, half of which were fatal (Gorby 1994). Immediate death occurs at 150 µg/m3. Extensive hemolysis ending in death follows 30 min-utes of exposure to 25 to 50 µg/m3 and less than 30 minutes after 100 µg/m3 (Ellenhorn 1997). The Office of Environmental Health Hazard Assessment (1999) describes acute toxicity after 1 hour exposition to 160 µg/m3. The odor threshold is 0.5 ppm or 1.6 mg/m3 at 25°C (New Jersey Hazard-ous Substances Fact Sheets 1993).

Feldmann et al. (1994) investigated gases from a domestic waste heap in Asslar (Germany) by LTGC-ICP-MS. They identified AsH3 via boiling point retention time correlation and semi-quantitatively measured total concentrations of volatile As (AsH3, DMA and TMA) to be between 16.2 and 48.5 µg/m3 relative to a liquid standard. Similar investigation on the gases from a thermo-philic and mesophilic digester showed concentrations of 6.4-25.3 µg/m3 total volatile As for the mesophilic and 16.1-30.4 µg/m3 for the thermophilic digestion (Feldmann and Kleinmann 1997). Arsine, however, was herein identified in traces only. Hirner et al. (1998) detected AsH3 in 2 sam-ples of geothermal gases in the Meager Creek area, British Columbia, by LTGC-ICP-MS but did not quantify it.

Another inorganic volatile As compound, just recently structurally analyzed by Kösters et al. (2003) is diarsine (As2H4) with a boiling point of 100°C. It has not been reported in environmental samples so far.

2.5.3.2 Monomethylarsine

Monomethylarsine (MMA, (CH3)AsH2) has a boiling point of +2°C (Lido 2003). Like AsH3 it can be produced by Methanobacterium formicicum (Michalke et al. 2000), as well as by Achromobac-ter sp., Aeromonas sp., Alcaligenes sp., Flavobacterium sp., Pseudomonas sp., and Nocardia sp. (Shariatpanahi et al. 1983). Concentrations detected above anaerobically incubated sewage sludge are comparable to AsH3 (0.68 ng/L, Michalke et al. 2000). MMA was not detected in the domestic waste heap gases by Feldmann et al. (1994), but in digester gases (Feldmann and Kleinmann 1997), and in 2 samples together with AsH3 in geothermal gases of British Columbia by Hirner et al. (1998). Another monomethylated volatile As compound with two As atoms, is monomethyl diar-sine (CH3)As2H3 (Kösters et al. 2003). It has not yet been identified in natural samples.

2.5.3.3 Dimethylarsine

Dimethylarsine (DMA, (CH3)2AsH) has a boiling point of +36°C (Lido 2003). Like AsH3 and MMA, it can be produced by Methanobacterium formicicum (Michalke et al. 2000) or by Escheria coli, Flavobacterium sp., Proteus sp., Pseudomonas sp. (Shariatpanahi et al. 1981), Achromobacter

Species 37

sp., Aeromonas sp., Alcaligenes sp., and Nocardia sp. (Shariatpanahi et al. 1983). Concentrations detected above anaerobically incubated sewage sludge are comparable to AsH3 and MMA (0.51 ng/L). DMA was detected in domestic waste heap gases by Feldmann et al. (1994), and in digester gases (Feldmann and Kleinmann 1997) but not in geothermal gases of British Columbia by Hirner et al. (1998).

2.5.3.4 Trimethylarsine

The completely methylated volatile As species, trimethylarsine (TMA, (CH3)3As) with a boiling point of +52°C (Lido 2003) is the most stable volatile As compound. It is also the most toxic form, except AsH3 because arsines gain toxicity with each added methyl group. Like the other volatile As compounds it can be produced by Methanobacterium formicicum (Michalke et al. 2000; Wicken-heiser et al. 1998) as well as by Pseudomonas sp. and Nocardia sp. (Shariatpanahi et al. 1981). Wickenheiser et al. (1998) pointed out that the predecessor for TMA production by Methanobacte-rium formicicum is DMAVA rather than TMAVO. Trimethylarsine was the predominant species (3.3 ng/L) above anaerobically incubated sewage sludge. According to these experiments, it is the only volatile As species formed by sulfate-reducing bacteria and the peptolytic bacteria Clostridium collagenovorans, Desulfovibrio gigas and Desulfovibrio vulgaris. According to Pickett et al. (1988), the most active anaerobic bacterium to volatilize trimethylarsine oxide to trimethylarsine from samples of river water, sea sediments, sewage sludge was Staphylococus aureus (208 nmol/min/g cells), the most active aerobic bacterium a marine pseudomonad (585 nmol/min/g cells).

Typical for all volatile As compounds is their distinct garlic-like smell. For TMA, the odor thresh-old is reported to be 2 ng/kg in a dilute aqueous solution (Yamauchi and Yamamura 1984b). Chal-lenger (1945) described this garlic-like odor from the breath of animals after being injected with DMAA. As mentioned before, TMA is the theoretical volatile end product of the As biomethyla-tion pathway but has never been found in human breath (Feldmann et al. 1996).

The occurrence of all volatile arsenicals in natural systems is limited due to their reactivity towards oxygen which decreases with an increasing degree of methylation. For TMA, the most stable com-pound, Parris and Brinckman (1976) report a second-order rate constant for gas-phase oxidation by O2 of 10-6 M-1s-1. Compared to (CH3)3Sb with an oxidation rate of 9 orders of magnitude higher (2⋅103 M-1s-1, assumed half-life: 50 ms, Jenkins et al. 1998; Haas and Feldmann 2000 found a half-life of > 24 h!), TMA appears to be relatively stable and could travel considerable distances with-out undergoing any chemical changes. Experiments by Pantsar-Kallio and Korpela (2000) confirm that TMA is relatively stable in air. Within 9 days only 30% of 10 µg TMA were oxidized to TMAO. This slow rate of oxidation explains why in most environmental samples TMA is the pre-dominant or only detectable arsine. Mukai et al. (1986) found a significant predominance of TMA, with ratios of 0.15 -0.34 for DMA to TMA when sampling airborne particulate matter at two sites (a polluted and a rural area) in Japan. As expected, the samples showed seasonal variations with higher concentrations in summer and lower concentrations in winter, in accordance with tempera-

38 Arsenic

ture changes. TMA was also detected in domestic waste heap gases by Feldmann et al. (1994), and as predominant volatile As species in gases from thermophilic and mesophilic digestion (Feldmann and Kleinmann 1997) as well as in geothermal gases of British Columbia by Hirner et al. (1998).

2.5.3.5 Organic Sulfur Arsines

Little is know about methylated sulfur arsines. Cullen and Reimer (1989) state that the rate of ar-sine evolution from (CH3)2AsSR is lower than from (CH3)2AsO(OH), and that the arsines produced are (CH3)3As and (CH3)2AsH, no volatile S-As compounds. One volatile As-S compound (dimethy-larsenomercaptane, (CH3)2AsS(CH3)), has already been synthesized in the laboratory by Ashe and Ludwig (1986) and by Kösters et al. (2003). So far, this compound has never received any attention in environmental studies, and has not yet been identified in natural samples.

2.5.3.6 Inorganic and organic halogen arsines

Tesfalidet and Irgum (1988) showed that it is possible to volatilize inorganic As(III) chemically as inorganic trichloroarsine (AsCl3) by acidifying dissolved ultrapure As2O3 with an excess amount of HCl ( Figure 13) and purging it, even without adding any reducing agent (commonly NaBH4). The aim of this study was to decrease interferences from metal ions that often occur during hydride generation analysis following the reduction with NaBH4 (section 2.6.1.1). Trichloroarsine (also named arsenic butter or fuming liquid arsenic) has a boiling point of 130°C. According to Arcand (1957), volatile AsC13 is the only As species present in significant quantity in 12 M HCl. Tri-chloroarsine and dissolved As(OH)C12 are the predominant species in 9 M HC1, while in 5.9 M HCl, AsC13, As(OH)C12, As(OH)2Cl, As(OH)2

+, and H3AsO3 all exist in significant quantities.

Mester and Sturgeon (2001) studied the formation of volatile chloroarsines during hydride genera-tion and detected the organic volatiles monomethylchloroarsine (CH3)AsCl2 (b.p. 133°C) and di-methylchloroarsine (CH3)2AsCl (b.p. unknown) besides the inorganic AsCl3. They confirmed that inorganic AsCl3 can form in the absence of a reducing agent. The volatilization of mono- and di-methylchloroarsines, however, required a strong reducing agent. Hydride and halide generation processes are similar and coexistent, with the share of halides increasing with increasing HCl con-centrations in solution. Contact of arsine with HCl vapor also resulted in the formation of AsCl3. Finally, Mester and Sturgeon (2002) found volatile arsines in a model atmosphere above seawater and sediment. The authors assumed that because the sediment was previously γ-irradiated, bio-methylation and biohydride generation can be excluded and the volatile species have to be halides. The halides seem to form abiotically, with an increasing release with increasing temperature. Kil-lelea and Aldstadt (2002) report the detection of dimethyl chloroarsine in the gas phase over a col-lected freshwater sediment core obtained from central Green Bay, Lake Michigan near a former herbicide factory. No studies were found of chloroarsine occurrence in the environment.

Analytical speciation techniques 39

2.6 Analytical speciation techniques

As mentioned before, since 1836 when Marsh for the first time positively identified As, there was no shortage of analytical procedures for the determination of total As. The volatile trimethylarsine (CH3)3As was detected already in 1933. Due to its volatilization from a bacterial culture it could be detected on a silver nitrate filter paper reducing the AgNO3 to elemental Ag. Further speciation was not achieved until 1973 when Braman and Foreback (1973) introduced the first hydride generation technique that was capable of separating different inorganic and methylarsenic compounds even in small concentrations. An excellent overview of As speciation analytics is provided in Gong et al. (2002).

2.6.1 Arsenic separation

2.6.1.1 Hydride generation

Hydride generation (HG) is based on the reduction of dissolved species, typically with sodium borohydride (NaBH4), to volatile species (hydrides) in an acid medium. Dissolved pentavalent As species are initially reduced to trivalent As species, before the arsines are produced according to the following two reactions:

RnAs(O)(OH)3-n + H+ + BH4- ↔ RnAs(OH)3-n + H2O + BH3

RnAs(OH)3-n + (3-n)BH4- + (3-n)H+ ↔ RnAsH3-n ↑ + (3-n)BH3 + (3-n) H2O

Borane, generated during these reductions, hydrolyses into the intermediate BH3OH- in a first reac-tion step and boric acid and gaseous hydrogen in a second reaction. The overall reaction is: BH3 + 3 H2O ↔ H3BO3 + 3H2 ↑

The advantage of HG for analysis is isolating the analyte from the matrix in the aqueous phase by transferring it to the gas phase (headspace). This process eliminates interferences and enhances sensitivities 10 to 100-fold (Howard 1997).

Typical concentrations for the reductant NaBH4 are 0.01-2%. Higher concentrations lead to in-creased back pressure in reaction vials due to the formation of excess hydrogen. Oxidation rates of NaBH4 in acid solutions are 6⋅105 L/mol/s for the first reaction step and 1⋅107 L/mol/s for the sec-ond, compared to much lower reaction rates in alkaline solutions with 2.9⋅10-4 L/mol/s and 4.7⋅10-2 L/mol/s (Gardiner and Collat 1965). Thus, for long-term stability during storage, NaBH4 is pre-pared in NaOH. For rapid NaBH4 oxidation with metals reduction during HG, an acid medium is used. The NaOH concentration for storage has to be optimized according to the amount of acid used during HG (Yin et al. 1996). Less commonly, NaB(C2H5)4 is used as a reducing agent, even though it has the advantage that the produced ethylates are more stable than hydrides generated with NaBH4 (Craig 1989). Feldmann et al. (2004) state that NaB(C2H5)4 as well as the also com-monly used Grignards reagents (e.g., C5H11MgBr) do not yield quantitative As derivatization.

The efficiency of arsine generation from individual As species is strongly dependent on pH. It seems that for the reaction to proceed rapidly, the species have to be present fully protonated, as

40 Arsenic

uncharged complexes (Howard 1997; Pantsar-Kallio and Korpela 2000). Negatively charged com-plexes would repel the electrons necessary for the reduction of the central As atom. Diagrams of arsine generation as a function of pH vary between different reports because of different acids used and different acid to reductant ratios (Bermejo-Barrera et al. 1998; Pantsar-Kallio and Korpela 2000; Rüde 1996; Shraim et al. 1999). Figure 14 summarizes the data according to Howard (1997).

Figure 14 pH de-pendence of arsine hy-dride generation from solutions containing As(V), MMAA, DMAA, and As(III) respectively; vertical arrows indicate the pK1 values for each acid (As(V) = 2.2, MMAA = 3.2, DMAA = 6.2, As(III) = 9.2 (modified from Howard 1997)

Arsenite predominates as H3AsO3

0 up to a pH of 9.2 and reacts with the reducing agent just at slightly acid conditions. Arsine generation from MMAVA must be carried out at a pH < 4, As(V) even below pH 2. The most effective arsine generation in both cases is below pH 1. DMAVA shows a clear peak in relative response at a pH between 0.5-2. The decrease at more acid conditions might be owed to the formation of the cationic DMAVA species (CH3)2AsO(OH)-H+ mentioned in section 2.5.2.1. Upon additional H+-input during acidification an H2O molecule easily splits from this spe-cies. A general decrease in response for all four As species can also be observed at excess HCl concentrations (not shown in Figure 14), since the surplus H+ rapidly consumes the reductant that does not volatilize the As species anymore (Luna et al. 2000). According to experiments from Sur (1999) (not shown in Figure 14) TMAVO shows a decreasing signal intensity with increasing HCl concentration. Trivalent methylated arsines and methylated sulfur arsines (not shown in Figure 14) can be generated like As (III) at a pH of about 6 (Bright et al. 1996; Hasegawa et al. 1994).

Trying to illuminate the observed variations in arsine-generation efficiency, Pergantis et al. (1997) used deuterium-labeled reagents for HG, and investigated the products with mass spectrometry. They found that AsH3, generated from As(III) or As(V), was deuterated when using deuterium labeled NaBH4 (NaBD4), however, not when using deuterium labeled HCl (DCl). MMA showed deuterium signature both when using NaBD4 or DCl. After an initial predominance of (CH3)AsD2, the concentration of non-deuterated (CH3)AsH2 increased with time. This time-series trend indi-cated that arsines once generated may further react with acids present in the reaction mixture, thus allowing for deuterium-hydrogen exchange. The relative rates of deuterium-hydrogen exchange depend on the basicity of the arsines. Arsine with the lowest basicity shows no exchange reactions,

0

10

2 0

3 0

4 0

5 0

6 0

7 0

8 0

9 0

10 0

110

-1 0 1 2 3 4 5 6 7 8

pH

rela

tive

resp

onse

[%]

H3AsIIIO30

H3AsVO40

(CH3)2AsVOOH

(CH3)AsVO(OH)2

pK1 pK1 pK1

pK1


whereas the most basic arsine, DMA, easily protonates according to the following reaction: (CH3)2AsH + H+ ↔ (CH3)2AsH2

+. The water-soluble dimethylarsonium (CH3)2AsH2+ is unstable

and oxidizes rapidly to (CH3)2AsO(OH)-H+. The hydrogen isotope signature in DMA equals that of the acid, i.e. it is (CH3)2AsD when using DCl, and (CH3)2AsH when using HCl. According to Per-gantis et al. (1997) deuterium-hydrogen exchange was only detected in the aqueous phase, not in the gas phase with HCl vapor.

The presence of other metals may suppress hydride generation (Creed et al. 1996), probably by decomposing NaBH4 before it has reacted with the analyte (Agterdenbos and Bax 1986). Total As recoveries are generally less affected by metal interferences than As(III) recoveries. Data from McCleskey et al. (2003) show that only Sb(III) and Sb(V) interfere with total As determination when the molar ratios of Sb(III)/Astotal exceed 4 and Sb(V)/Astotal exceed 2. Poor As(III) recoveries because of hindered arsine generation or oxidation of As(III) were found when the following ratios were exceeded: Cd/As(III) > 800, Cr(VI)/As(III) > 2, Cu(II)/As(III) > 120, Fe(III)/As(III) > 70, Sb(III)/As(III) > 3, Sb(V)/As(III) > 12, and Se(IV)/As(III) > 1. High Fe(II) concentrations (Fe(II):Fe(III) > 2) inhibited the As(III) oxidation by the oxidizing agent Fe(III). Poor As(III) re-coveries in the presence of Fe(II):Fe(III) ratios > 2 occurred only at Fe(III) concentrations of more than 10 mg/L. The elements Fe(II), Al, Cr(III), Co, Pb, Mn, Ni, and Se(VI) did not lead to signifi-cant decreases in arsine generation (McCleskey et al. 2003). In a 2 µg/L As(V) solution, Yin et al. (1996) observed significant interferences only from a 150-fold excess of Se(IV) and Bi(III), slight signal depressions from Fe(III) (10 mg/L) and Ni (10 mg/L), but none from Pb(II), Ag(I), Al(III), Ga(II). In contrast to McCleskey et al. (2001, 2003), they found no interference from Cu(II) up to 10 mg/L. In experiments conducted by Jamoussi et al. (1996) 1 mg/L Au(III) and Pd(II), 2 mg/L Ni (II), 25 mg/L Cu(II), and 50 mg/L Co(II) severely suppressed the As signal (recovery < 80%). Lit-tle interference was found for Pb(II) with a recovery of more than 90% for 100 mg/L Pb(II).

Interfering metal cations can be removed by a cation exchange resin without alteration of the origi-nal As(III)/Astotal ratio (McCleskey et al. 2003). Pre-reduction of As with cysteine also yields a greater resistance to interferences from metal ions by forming stable ligand-metal complexes with the metal ions as well as more uniform sensitivities for the four main As species (Howard 1997; Jamoussi et al. 1996; Yin et al. 1996). The latter is probably caused by the lower acid concentra-tions when working with pre-reduction, generating less hydrogen which is advantageous for the atomization of As (Yin et al. 1996). Similarly, Yano et al. (2000) found much better recoveries for As(V) using potassium iodide and microwave treatment for pre-reduction, working at lower HCl and NaBH4 concentrations.

For total As determination by HG, the pH dependent difference in signal response of the different As species poses a problem, since a compromise for the pH has to be found for all species or pre-reduction steps have to be performed. On the other hand, these differences can be used to achieve simple As speciation by stepwise HG at pH 6, yielding trivalent As species, and consequently at pH 1, for the sum of tri- and pentavalent As species. Pentavalent species are then calculated from the difference. Further speciation between inorganic, mono-, di- and trimethylated As is mostly done

42 Arsenic

by subsequent cryogenic trapping and sequential release to a detector according to boiling point differences (section 2.6.2).

However, by laborious optimization of acid, reductant, and pre-reductant concentrations as well as reagents contact time, it is possible to develop a selective reduction scheme for the identification of As(III), As(V), MMAA, and DMAA. Bermejo-Barrera et al. (1998) used a citric acid / sodium hydroxide buffer at pH 5 and 0.2% NaBH4 to volatilize As(III) alone, 0.14 M acetic acid and 0.2% NaBH4 for As(III) + DMAA, 4 M HCl + 2% NaBH4 for As(III) + As(V), and 0.02 M thioglycollic acid + 1 % NaBH4 for HG of all As species together. Shraim et al. (1999) used 0.01 M HCl, 2% NaBH4, 5% L-cysteine and a contact time <10 min for total As, 1 M HCl, 0.3-0.6% NaBH4, 4% L-cysteine, < 5 min for DMAA, 4-6 M HCl, 0.05% NaBH4, no L-cysteine for As(III), 4 M HCl, 0.03% NaBH4, 0.4% L-cysteine, 30 min for MMA determination. As mentioned before, reports about efficiency of HG from individual As species differ, being highly sensitive towards type and amount of reagents. Shraim et al. (1999) also discuss some of these contradictions. The successful separation of As(III) + As(V) at 4 M HCl by Bermejo-Barrera et al. (1998) is based upon their ex-periments; no MMA was volatilized at any HCl concentration. This finding contradicts to many other reports (e.g., Howard 1997; Shraim et al. 1999). Furthermore, Shraim´s separation of DMAA is based on their findings that DMAA shows the highest signal response at low pH and low NaBH4 which was not confirmed by Howard (1997). Determination of inorganic and organic As species by hydride generation is possible but still far from being a standardized routine application.

The HG techniques have some further restrictions. First, in contrast to the inorganic and organic As species described above, several organoarsenicals are not reducible with borohydride, e.g., arseno-betaine, arsenocholine, or arsenosugars. Those refractory or “hidden” arsenicals have to be pre-treated e.g., by alkaline digestion (de Bettencourt and Andreae 1991), microwave-assisted oxida-tion (Le et al. 1993; Santosa 2001; Sur 1999), UV photooxidation with K2S2O8 (Howard and Hunt 1993; Sur 1999; van Elteren and Slejkovec 1997) or a combination of all three (Cabon and Cabon 2000; Yano et al. 2000) to become convertible to arsines. While the hybrid ion arsenobetaine is more resistant to oxidation than the cationic arsenocholine species, it is vice versa for alkaline di-gestion (Sur 1999).

The HG techniques include the simplified assumption that AsH3 only forms from inorganic As, MMA only from monomethylated species, DMA only from dimethylated species, TMA only from TMAVO. Transmethylation reactions during volatilization and later on in the gas phase are hereby neglected. Already Talmi and Bostick (1975) found that the reduction of MMAA resulted in the partial production of AsH3, DMA, and TMA in addition to the expected MMA. Hydride generation on DMAA produced small amounts of MMA. According to experiments from Pantsar-Kallio and Korpela (2000), AsH3 and MMA are not converted into other gaseous species during preparation or storage. For DMAA they found that at pH 1 to 2, about 90% of DMAA are volatilized as DMA, a small fraction, however, as TMA and MMA. This fraction increases with decreasing pH. At pH lower than -0.3 all DMAVA is converted to TMA.


Furthermore, HG of MMAVA and DMAVA may not only yield the simple methylated arsines, but also more complex forms as methylated chloroarsines in the presence of excess Cl in solution or in the gas phase (section 2.5.2.6). Van Elteren et al. (1994) e.g., report that in their experiments only 85% of the originally present DMAA was converted to DMA. The remaining 15% showed a higher affinity to the gas chromatographic column, indicating a compound with a higher boiling point. The compound was unknown to the authors and could have been a dimethylated chloroarsine. Haas and Feldmann (2000) found a similar phenomenon of formation of various hydrides from one precursor during HG for Sb. They detected SbH3, (CH3)SbH2, (CH3)2SbH, (CH3)3Sb only from (CH3)3SbCl2 as precursor due to rearrangement and demethylation reactions. Bergmann and Glindemann (1997) also detected other volatile As species besides the expected AsH3 after HG of an As(V) solution. The concentrations of these additional arsines increased with increasing concentration of dissolved As(V). The authors explain this trend by the formation of oligomers from the monohydrides or the formation of methylated arsines from contaminations in the reducing agent NaBH4. Re-dissolution of the generated hydrides might further increase losses and uncertainties. Using flow injection HG (FI-HG) instead of batch-wise operation, first developed and applied by Aström (1982) and Ya-mamoto et al. (1985), will decrease re-dissolution and interactions in the headspace. However, there might still be significant errors in simply concluding from the volatile species detected during HG to the dissolved species (Feldmann et al. 2004).

Much more reliable than HG alone is a combination of HG with a preceding separation step, like chromatographic separation by high performance liquid chromatography (HPLC). For chromatog-raphy the sample is introduced unaltered, differences in boiling points yield different retention times and an easy species separation. Subsequent HG is used for improved detection limits.

2.6.1.2 HPLC

For As speciation via separation by high performance liquid chromatography (HPLC), the tech-nique of ion pair, ion exchange and size exclusion chromatography are used. The most commonly used pairing cation is tetrabutyl ammonium (elution order As(III) > DMAVA > MMAVA > As(V) in a pH between 5 and 7). Arsenite and arsenobetaine co-elute in this pH range. To separate both species, anionic-pairing conditions and a higher mobile-phase pH (pH >9) have to be used so that As(III) is slightly retarded and while the hybrid ion arsenobetaine passes through the column (Thomas and Sniatecki 1995). Using a reversed-phase column and a mobile phase of 5 mM tetrabu-tyl ammonium hydroxide (TEAH), 3 mM malonic acid, and 5% methanol at pH 5.8, MMAIIIA and DMAIIIA were also determined by Le et al. (2000a, 2000b). Le and Ma (1997) used a reversed-phase column and a mixed ion-pair mobile phase of 10 mM hexanesulfonate and 1 mM TEAH to separate As(III), As(V), MMAVA, DMAVA, arsenobetaine, arsenocholine, and TETRA. Londes-borough et al. (1999) separated optimally those species plus TMAVO with a 0.05 mM benzenedi-sulfonic acid as eluent modifier. They encountered, however, problems resolving As(III) and MMAA as well as identifying TMAVO, TETRA, and arsenocholine in natural samples because of matrix effects.

44 Arsenic

Different research groups developed methods for the separation of the four main As species (As(III), As(V), MMAVA, DMAVA) by gradient elution on anion exchangers, mainly styrol-divinylbenzole copolymers. For the mobile phase, Stummeyer et al. (1996) and Bednar et al. (2004) used 2.5 mM to 50 mM, and 38 mM to 75 mM phosphate buffers respectively, Lindemann et al. (1999) used 2 mM ammonium hydrogen carbonate, and 2.2 to 45 mM tartaric acid at a pH of 8.2, whereas Bissen and Frimmel (2000) used 2 mM and 50 mM NaOH. Using a boric acid buffer (10 mmol and 50 mmol Na2B4O7⋅10H2O in 20 and 30 mmol NaCl) at pH 9 instead of the weaker phos-phorous buffer, Sur (1999) could additionally separate TMAVO, however with the drawback of relatively long analysis times. By gradient elution with nitric acid (2.5 mM and 50 mm) Bednar et al. (2004) were able to resolve roxarsone (3-nitro-4-hydroxyphenylarsonic acid) besides the four main As species. Cation exchange chromatography was successfully used to separate arsenobe-taine, arsenocholine, TMAVO, and TETRA (Ackely et al. 1999). Feldmann et al. (2004) also achieved separation of arsenosugars and their metabolites via cation exchange.

In the ion exclusion mode, a carboxylated methacrylate resin separates As(V), MMAVA, DMAVA, As(III), and arsenobetaine. However, differentiation of TMAO and arsenocholine was not achieved (Taniguchi et al. 1999). McSheehy et al. (2000) used size exclusion to separate arsenosugars from other As species.

2.6.1.3 Gas chromatography

Separation of volatile species or originally dissolved species after HG can be achieved by gas chromatography, often combined with cryofocusing, i.e. the pre-concentration of the analyte in a cooled trap (section 2.7.4.3). Different hydride species can be separated by sequential release due to boiling point differences. Various columns have been used for separation, e.g., Chromosorb coated with 10% methyl silicon SP-2100 (Amouroux et al. 1998; Feldmann et al. 1994; Grüter et al. 2000; Hirner et al. 2000), a DB624 column with 6% cyanopropylphenyl and 94% dimethyl-polysiloxane (medium polarity, Prohaska et al. 1999), or a SPB-5 with 5% phenyl and 95% di-methylpolysiloxane (low polarity, Mester and Sturgeon 2001).

2.6.1.4 Capillary zone electrophoresis

Another way of separation is capillary zone electrophoresis (CZE), used by Schlegel et al. (1996) for the separation of As(V), As(III), DMAA, AsVF6, p-aminobenzenearsonate, and phenylarseo-nate, as well as by Greschonig et al. (1998) for As(III), As(V), DMAA, diphenylarsinic acid, methanearsonic acid, phenyl- and p-aminophenyl arsonic acid, phenylarsineoxide and phenar-sazinic acid and by Michalke and Schramel (1997) for As(III), As(V), DMAVA, MMAVA, arseno-betaine and arsenocholine. Problems occurred in some studies with As(V). At alkaline pH it could not be analyzed, because its electrophoretic mobility was higher than the velocity of the electroos-motic flow, at acidic pH analysis time was long. In general, CZE proved to have a worse sensitivity compared to HPLC with a poor resolution especially in complex matrices (Gong et al. 2002). To


improve the limits of detection in CZE, pre-concentration by stacking larger sample volumes and removal of the sample matrix were performed, normally at negative polarity or by adding a modi-fier, or, as proposed by Vanifatova et al. (1999), by using a polymer-coated capillary to decrease the zeta potential at the capillary walls which diminishes the electroosmotic flow.

2.6.2 Arsenic detection

Arsenic detection after HG can be done e.g., by coupled AAS (atomic absorption spectrometry), AFS (atomic fluorescence spectrometry that shows a very high sensitivity in the absence of light scattering and background interference from the sample matrix after HG), ICP-AES (inductively coupled plasma atomic emission spectrometry), or ICP-MS (inductively coupled plasma mass spec-trometry). Bermejo-Barrera et al. (1998) describe electrothermal AAS (ET-AAS) with in situ pre-concentration of arsines on coated graphite tubes as superior to conventional HG-AAS. Testing Ir-, W-, Zr-coatings, they found the best analytical performance with detection limits of around 50 ng/L for Zr coatings.

Another possibility that is used both for volatile species as well as for originally dissolved species after HG, is cryofocusing (section 2.7.4.3) and subsequent injection into AAS (Cabon and Cabon 2000; Howard 1997; Van Elteren et al. 1994) or GC (LT-GC, low temperature GC). The advantage is that different hydride species can be detected by sequential release from the cold trap based on boiling point differences. Unknown compounds can be detected by boiling point – retention time relationships (Feldmann et al. 1994; Grüter et al. 2000), or, in the case of GC, by additional detec-tors, e.g., ICP-MS (Amouroux et al. 1998; Feldmann et al. 1996; Feldmann 1997; Hirner et al. 2000; Prohaska et al. 1999). Especially for As, however, there is the problem of validating ICP-MS signals caused by interferences of 35Cl and 40Ar with the mono-isotopic 75As (Pécheyran et al. 2000). In the absence of known standards, structural information and species identification can only be gained using GC-MS (Cullen et al. 1995; Feldmann and Kleinmann 1997; Kaise et al. 1997; Maillefer et al. 2003; Pantsar-Kallio and Korpela 2000). Electrospray ionization (ESI) MS, possibly combined with tandem mass spectrometry (ESI-MS/MS), is also well suited for As speci-ation, especially of arsenosugars (Gallagher et al. 2001a; McSheehy et al. 2000; Pergantis et al. 2000; Shimizu et al. 1999).

HPLC or CZE can further be coupled with • UV- or conductivity detection • HG-AAS (Carrero et al. 2001; Howard and Hunt 1993; Shraim et al. 2000; Sur 1999) • HG-AFS (Le and Ma 1998; Le et al. 2000a; Ma and Le 1998; van Elteren and Slejkovec 1997) • HG-ICP-AES (Gettar et al. 2000; Rauret et al. 1991) • HG-ICP-MS (Prohaska et al. 1999) • IPC-AES (Chausseau et al. 2000; Ebdon et al. 1987; Weeger et al. 1999) or • ICP-MS (Guerin et al. 2000; Lindemann et al. 1999; Londesborough et al. 1999; Saeki et al.

2000).

46 Arsenic

Gómez et al. (1997) propose anion cartridge pre-concentrators for improvement of the detection limits of HPLC-HG-AAS (0.1 - 0.6 µg/L). Often, the post column reaction loop in a HPLC-HG-AAS setup is heated to increase signal intensity, especially for the methylated arsenicals. Stum-meyer et a. (1996) found a 20% and 40% increase in signal intensity for MMAA and DMAA re-spectively, by heating the reaction loop to 100°C compared to reduction at room temperature. De-tection limit was 0.5 µg/L.

Detection limits of HPLC-HG-AFS are in the sub-microgram per liter range, comparable to those achieved with HPLC-ICP-MS using pneumatic nebulization. Le et al. (2000a) successfully sepa-rated As(III), As(V), MMAIIIA, MMAVA, DMAIIIA, DMAVA by HPLC-HG-AFS with detection limits of 0.5-2 µg/L without further sample pre-treatment. The ICP techniques are especially inter-esting for multi-element studies. Coupling HPLC and ICP-AES is straightforward because the flow rates in both cases are similar, typically about 1 mL/min. However, ICP-AES is mostly used for samples with higher As concentrations because it has poor sensitivity. For higher sensitivity either a HG step is established between HPLC and ICP-AES or ICP-MS is used with about 100 to 1000 times higher sensitivities compared to ICP-AES, provided corrections for Cl-Ar interferences are made.

Less common techniques include a gold ultramicroelectrode array proposed by Feeney and Koun-aves (2000) that responds with high sensitivity to As(III) using square-wave anodic stripping volt-ammetry, or biosensors, such as the biosensing microbe Pseudomonas fluorescens OS8 (Petänen et al. 2001) or the GFP (green fluorescent protein) reporter gene biosensor that is fused to the ars op-eron protein ArsR for arsenite repression (section 2.5.1.2) and produces GFP in response to As(III) (Roberto et al. 2002).

2.6.3 Arsenic quantification

While the quantification of dissolved As species is relatively simple via aqueous standard solutions, the quantification of the gaseous As species is more difficult since few commercial standard gas mixtures and no reference material exist. High pressure standards of AsH3, MMA, DMA, and TMA were synthesized e.g., by Bergmann and Glindemann (1997). The pure As substances were evapo-rated in evacuated high-pressure stainless steel gas cylinders and stored in a He-atmosphere.

Calibration is often performed externally with gaseous standards produced by HG and transferred offline to a GC (Prohaska et al. 1999). Feldmann (1995, 1997) alternatively developed an online calibration method for GC-ICP-MS. The dry aerosol of the gas sample from the GC output is mixed in a T-piece or in a high-solid torch with an aqueous standard solution containing Rh as continuous internal standard to monitor the plasma stability and identify irregular transient signals caused by the presence of CH4 or CO2. The aqueous standard is introduced as a wet aerosol by nebulization. Nebulized water is also added to the dry aerosol from the GC outlet in order to obtain the same response for volatile and aqueous species during their simultaneous introduction into the plasma. Otherwise the different water content in the plasma would cause significant differences in

Field sampling 47

electron density, ion energy, and ionization temperature. While the external calibration with HG requires multi-species standards (AsH3, MMA, DMA, etc.), the Feldmann method is based on the species independent calibration with only one As-specific response curve. An error of ± 35% can be expected for the determination of the mass of the internal Rh standard.

2.7 Field sampling

2.7.1 Water sample stability and preservation

Several papers on the stability of As species over time show that generally rapid oxidation of As(III) to As(V) occurs in natural water samples without preservation. Hall et al. (1999) report a complete reduction of As(V) to As(III) within a few days storage at room temperature for samples of Ottawa river water and deionized water. Such a reduction can occur during storage when the water sample contains dissolved organic carbon as food source for microorganisms and turns an-oxic because it is sealed. In the case of Hall et al. (1999) the microorganisms most likely came from organic resins used in the deionization cartridges. Seyler and Martin (1989) found no changes in the As(III) concentrations at all after 10 days storage at 4°C and a pH of 7. McCleskey et al. (2004) detected such an unchanged As redox species distribution only in distilled, deionized, and otherwise uncontaminated water (16 days at pH 3.7 under light and dark conditions). For the or-ganic arsenicals Larsen et al. (1993) report that DMAVA, MMAVA, and arsenobetaine are rela-tively stable. Jokai et al. (1998) stored solutions with 0.01-10 ng/mL MMAA and DMAA over 5-6 months at room temperature without any significant losses. Trivalent methylated compounds, espe-cially DMAIIIA, however, oxidize rapidly (Gong et al. 2001; section 2.5.2.3).

As discussed in section 2.5.1.2, oxidation of As(III) to As(V) or reduction of As(V) to As(III) are mainly microbial processes but inorganic reduction can occur in the presence of strong reducing agents such as H2S and S2O3, inorganic oxidation in the presence of metal catalysts such as MnO2 or strong oxidizing agents such as Fe(III) especially when catalyzed by light. Species preservation methods therefore must include filtration of colloids and microorganisms, addition of a reagent to prevent dissolved Fe and Mn oxidation and precipitation, and cool and dark storage (McCleskey et al. 2004). Creating an inert N2 atmosphere shows no improvement for species preservation (McCleskey et al. 2004).

Filtration shall be done with 0.1 µm (McCleskey et al. 2004). For storage, Palacios et al. (1997) recommend freezing as the best way. However, freezing promotes the formation of iron hydroxides leading to oxidation, sorption, and co-precipitation of As. Lindemann et al. (2000) and Daus et al. (2002) found the worst storage results at -20°C, the best at 3°C, similar to McCleskey et al. (2004) at 4°C. Various reagents have been tested for stabilization. Filtration alone without adding any stabilizing agent does not inhibit algal growth (McCleskey et al. 2004). McCleskey et al. (2004) recommend acidification with HCl to pH 2 as stabilization. Dark storage is crucial because highly reactive dichlororadicals can form under the influence of light and lead to rapid oxidation. Nitric

48 Arsenic

acid can undergo photochemical reaction and should thus not be used for redox species preserva-tion (Fanning 2000). Sulfuric acid works for the preservation of As redox species (Bednar et al. 2002) but it is difficult to purify and forms precipitates with many other metals such as BaSO4. Daus et al. (2002) used 0.01 mol/L phosphoric acid because they found that using the strong acid HCl or the weak nitrilotriacetic acid as preservatives led to an overestimation of As(V) by dissolv-ing colloidal soluble adsorbates containing arsenate and by promoting ferric ion formation. McCleskey et al. (2004) point out that Daus et al. (2002) did not exclude light from their experi-ments and that thus As(III) oxidation is to be expected. The use of phosphoric acid has the disad-vantage that it may encourage microbial growth and may lead to supersaturation with respect to metal phosphates and sorption or co-precipitation of As or other redox sensitive elements. Edwards et al. (1998) used ascorbic acid and HCl to preserve the As(III) and As(V) concentrations for 28 days in spiked synthetic water, however in natural samples 25% of the As(III) was converted to As(V) within 8 days despite of the preservation measures. Using an optimized method of filtration with 0.1 µm, acidification with HCl to pH 2, and storage at 4°C in opaque bottles the As(III) to As(V) ratio was found to be stable for more than 15 months in surface and groundwaters (McCleskey and Nordstrom 2002; McCleskey et al. 2003, 2004).

An excess of Fe(II) was used successfully by Borho and Wilderer (1997) to inhibit As(III) oxida-tion in Fe(III) rich samples. McCleskey et al. (2004) showed that addition of Fe(II) and SO4

2- slows down the As(III) oxidation by Fe(III) in light experiments and stops it for 71 days in dark experi-ments. The Fe(II) competes with As(III) for dichloro- and hydroxyl radicals, SO4

2- complexes with Fe(III) and the FeSO4

+ complex sorbs photons, thus binding free radials.

If detection is to be performed by ICP-MS McCleskey et al. (2004) and Bednar et al. (2002, 2004) recommend ethylenediaminetetraacetic acid (EDTA) as preservative because of the molecular in-terference of 35Cl from HCl and 40Ar from the carrier gas on the monoisotopic 75As. The EDTA chelates Fe and thus prevents photochemical oxidation and precipitation. Gallagher et al. (2001b) also report that As(III) concentrations were stable for 16 days after addition of EDTA. The disad-vantage is that relatively high concentrations of EDTA, partly exceeding its solubility, are needed in Fe rich waters. Manganese, a similarly efficient catalyst for As(III) oxidation, is not chelated at all. A further disadvantage is that metals sequestered by EDTA can not be removed by cation ex-change as described in section 2.6.1.1 (McCleskey et al. 2003) and might interfere with arsine gen-eration.

For human urine Feldmann et al. (1999b) report a significant reduction of As(V) to As(III) as well as a demethylation of MMAA and DMAA to As(III) with strong acidification (0.1 mol/L HCl). They found no improvement of species recovery with other preservatives tested (sodium azide, benzyltrimethylammonium chloride, benzoic acid, cetylpyridinium chloride, methanol) and state that samples can best be kept without any additives at 4°C or -20°C for 2 months. The problems of storage at -20°C were discussed above.

Field sampling 49

2.7.2 On-site species separation for water samples

To exclude potential losses during storage and transport, several on-site species separation tech-niques have been developed. Van Elteren et al. (1994) propose on-site sequential coprecipitation of As(III) with carbamate and As(V) with molybdate. The precipitates are filtered under vacuum on a 0.45 µm filter and dissolved with HCl, HNO3 or H2SO4 upon return to the laboratory. Hasegawa et al. (1994) developed another method that was successfully applied by Sohrin et al. (1997) for the separation of six As species. All As(III) species (i.e. As(III), MMAIIIA, DMAIIIA) are extracted on site with diethylammonium diethyldithiocarbamate (DDDC) into carbon tetrachloride. In the labo-ratory this aliquot is back-extracted into 0.1 M NaOH, including several clean-up steps to remove the remaining DDDC, and analyzed by HG-AAS. The sum of As(III) and As(V) species are ana-lyzed from a different aliquot filtered with 0.45 µm upon collection and acidified to pH 2. Finally, the As(V) species (As(V), MMAVA, DMAVA) are obtained as the difference between total As and As(III). One problem with this method is that trivalent As-S-compounds can not be identified or distinguished from MMAIIIA, DMAIIIA which are extracted as dithioldicarbamate complexes. An-other drawback is the relatively complicated and time consuming back-extraction in the laboratory.

Ion exchange on solid phases (SPE = solid-phase extraction) has been used as a rather simple, in-expensive, and sensitive method for As(III) and As(V) on-site separation (Ficklin 1983; Kim 2001). The species separation is based on the different dissociation constants of As(III) and As(V) as shown in Figure 4 and Figure 6. The method was further improved by Le et al. (2000c) and Yal-cin and Le (2001) for the additional separation of the methylated arsenicals. A filtered water sam-ple is passed through a resin-based strong cation exchanger, on which DMAA is sorbed, followed by a silica-based strong anion exchanger that retains MMAA and As(V). Arsenite as uncharged complex passes through both exchange cartridges and remains in the solution. The sample has to be filtered to avoid precipitation of Fe, Mn, or Al-hydroxides at the head of the cartridge that would provide additional sorption sites (Bednar et al. 2004). The sequence of cation exchanger mounted on top of the anion exchanger is important because DMAA that predominates as uncharged com-plex at pH > 1.6 ( Figure 9) shows non-specific sorption both on the cation and anion exchanger, whereas MMAA and As(V), as negatively charged complexes, pass through the cation exchange cartridge and sorb on the anion exchanger only. Concomitant anions, like sulfate, carbonate, or chloride can compete with As species for exchange sites if present in high concentrations, such as sulfate in an acid mine drainage (AMD) with concentrations in the g/L range (Bednar et al. 2004).

In the laboratory, the cation exchange cartridge is eluted with 1 M HCl to mobilize the sorbed DMAA fraction. The anion exchange cartridge is first rinsed with 60 mM acetic acid at a pH of approximately 4 to mobilize MMAA with a pK1 value of 3.6 ( Figure 8). At this pH, As(V) is still retained as negatively charged complex. It only desorbs after treatment with 1 M HCl at a pH of approximately 1 (pK1 = 2.2; Figure 4). Each eluent fraction is analyzed separately by HG-AAS with a detection limit of 0.05 µg/L. Sealed and stored in the refrigerator at 4°C, the As species were found to be stable for up to 4 weeks on the ion exchangers. Instead of pre-conditioning the ex-changer cartridges with 50% methanol and deionized water (Le et al. 2000c), Bednar et al. (2004)

50 Arsenic

suggest converting the resin from the chloride form to the acetate form by washing it with 1.7 M acetic acid to achieve higher buffering capacity and a better long-term stability (recovery after 47 days of storage 99% for the acetate form compared to 85% for the chloride form).

2.7.3 Gas sampling without pre-concentration

Gas containers, or gas bags are used to take whole air gas samples without pre-concentration. Schweigkofler and Niessner (1999) used evacuated stainless steel gas containers with electro-polished internal surfaces for chemical inertness described in the EPA canister method TO-14 (Fal-ter et al. 1999) for storing volatile organic carbons (VOC). The recovery for the non-polar com-pounds was more than 85% and significant losses occurred for polar compounds only, e.g., propa-nethiol, and butanethiol. No applications for volatile metallics were found in the literature.

Gas bags of Teflon or Tedlar have the advantage of an easier handling compared to gas canisters. A variety of gas volumes (1-100 L) can be taken, whereas canister volume is limited to a few liters unless the sampled air is pressurized. Applications of gas bags are described by Feldmann et al. (1996), Feldmann and Cullen (1997), and Hirner et al. (1998). Haas and Feldmann (2000) investi-gated the stability of volatile As, Sb, and Sn in Tedlar bags. Surprisingly, the most stable com-pound was AsH3 with a recovery rate of 75% after 8 weeks storage at 20°C and an estimated half-life of 3-4 months. The recovery rate of TMA on the other hand was only 75% after 24 hours of storage and 15% after 5 weeks.

Commonly, losses in gas bags occur due to sorption effects. Andino and Butler (1991) found that water vapor sorbs on the Tedlar walls and subsequent partitioning of VOCs into the absorbed water can occur. Wang et al. (1996) found for the VOCs trichloroethylene, dichloroethane, and toluene that commercial Tedlar bags showed significant sorption at the stainless steel hose valves or septa. They exclude losses due to leakage because the decrease was exponential and typical for a finite number of available sorption sites. Also surface to volume ratios have an effect on sorption losses, with larger volumes showing less sorption losses.

Haas and Feldmann (2000), however, concluded that for their experiments with concentrations between 0.3 to 10 ng/L sorption or condensation on the walls of the bag were a minor problem, especially when working with higher concentrations. Significant losses are attributed to immobili-zation of the volatile arsines due to oxidation and demethylation. The differences in stability found between AsH3 and TMA are explained by the different dissociation energies for As-H (352 kJ/mol for AsH3) and As-C (averaged dissociation energy 262 kJ/mol for TMA) bonds. These findings contrast with experiments from Pantsar-Kallio and Korpela (2000) and Parris and Brinckman (1976) that TMA is more stable than AsH3 (section 2.5.3.1, 2.5.3.4). The dissociation energy for As-H cited in Haas and Feldmann (2000; original from Thayer 1995) seems too high, compared to 229 kJ/mol in Elsenbroich and Salzer (1992). Independent of these unresolved differences, Tedlar bags can be successfully applied for gas sampling, if analysis is completed after a maximum of 8 hours storage, where all arsines could be recovered to approximately 100%.

Field sampling 51

Simple storage of gas samples, without any pre-concentration steps has several advantages. In con-trast to trapping techniques described below there is no potential danger of analyte loss due to breakthrough. The whole gas composition can be determined, whereas trapping techniques are mostly optimized for selected compounds. Last but not least, multiple analyses can be performed from one sample.

2.7.4 Gas sampling with pre-concentration

The major challenge for sampling volatile metallics are their low concentrations, and possible inter-ferences with other gas components (typically CO2, CH4, H2S). Selective trapping with pre-concentration is therefore the preferred choice. Trapping gas samples over a certain period of time, so-called integrative sampling, leads to a pre-concentration of the volatile metallics, and smoothes out short term temporal noise. Liquid and solid sorbents are used as trapping media.

2.7.4.1 Liquid sorbents

Commonly used liquid sorbents for volatile metallics are oxidizing reagents, e.g., NaOCl, H2O2, Na2S2O8, or Na2O2 (Klusman 1991). Ca(OCl)2 is more reactive than NaOCl, and more soluble in water. However, because of its instability as solid substance and in solution Ca(OCl)2 is generally less suited for analysis than NaOCl. For As also the solubility of Ca3(AsO4)2 or CaHAsO4·H2O is a limiting factor (section 2.5.1.1; Bothe and Brown 1999). The maximum amount of As soluble in a 50 mL 1M Ca(OCl)2 solution is 12.2 mg As at 25°C. For sulfur containing gases, only 0.12 mg sulfate could be collected without precipitating calcium sulfate, and probably decreasing As con-centrations by co-precipitation or sorption (Klusman 1991).

The use of liquid sorbents is well suited as a fast screening method, since a large variety of volatile metallics can be trapped efficiently, stabilized up to several weeks, and analyzed by multi-element analysis. However, the oxidation results in a complete destruction of species information, the indi-vidual arsines will all be detected as the sum of volatile As. In contrast to liquid sorbents, the solid sorbent methods described in the following section are species-conserving.

2.7.4.2 Solid sorbents

Two types of species-conserving solid sorbents are presented in the following: glass tubes filled with sorptive material comparable to small (max. 10 cm length) chromatographic columns (sorp-tion tubes), and microfibers coated with sorptive material, the so-called SPME (solid-phase micro extraction) technique. In both cases, the sorption of volatile metallics on the sorbent is followed by thermal desorption and capillary gas chromatographic analysis.

Synthetic porous polymers (Tenax, Porapak, Chromosorb), carbon molecular sieves (Carbosieve, Carboxen) or non-porous graphitized carbon (Carbotrap) are commonly used as sorptive materials with a large surface area for quantitative sorption of volatile compounds ( Table 4). Especially

52 Arsenic

Tenax finds widespread application for environmental sampling due to its wide range of possible sorbents, relative hydrophobicity, reproducible behavior during thermal desorption, and low chro-matographic background noise.

Table 4 Different sorption materials for sorption tubes and their properties (Supelco catalogue information; elemental composition according to Knobloch and Engewald 1995)

Elemental composition [%]

sorp

tion

mat

eria

l

type

desc

riptio

n

max

. tem

p. [°

C]

spec

ific

area

[m2 /g

]

sam

plin

g ra

nge

brea

kthr

ough

vol

ume

[H2O

30°

C, m

L/g]

C

H

N S O

Tenax TA porous polymer

2.6 diphenyl-p-phenylene ether 300 35 C6-C14 32 88.5 5.0 --- --- 6.3

Chromo-sorb 105

porous polymer

polyaromatic, for com-pounds with a boiling

point ca. 200°C 225 650 C5-C12 --- --- --- --- --- ---

Chromo-sorb 106

porous polymer

cross-linked polystyrene, low boiling compounds, gases, C2-C5 alcohols

250 750 C2-C5 ---

--- --- --- --- ---

Carbosieve S III

carbon molecular

sieve

small molecules, hydro-philic 400 820 C2-C5 111 96.8 0.5 <1 --- ---

Carboxen 564

carbon molecular

sieve

C2-C5 volatile organic carbons 400 426 C2-C5 --- 82.1 0.6 <1 6.3 ---

Carbotrap graphitized

carbon black

C4-C5 to polychlori-nated biphenyles, larger

molecules 350 100 C5-C9 0 99.6 --- 0.2 --- ---

Jenkins et al. (1998) used Tenax TA to trap (CH3)3Sb and encountered oxidation on the sorption material under aerobic conditions. The oxidation product was less volatile, tightly bound to the sorptive, and not released during thermal desorption, but only after extraction with nitric acid. Transferring the analyte to the sorption material under anaerobic conditions enabled reversible sorption. Grüter et al. (2000) detected the same irreversible sorption of Sb and losses of over 50% with different tin and arsenic species during sorption studies on Tenax at room temperature (18 °C). Other potential problems using Tenax sorption tubes are degradation by exposure to oxidants like NOx, O3, H2O2, and OH radicals (Kleno et al. 2002; Pellizzari and Krost 1984) and alteration by irradiation with sunlight (Calogirou et al. 1996; Peters et al. 1994), producing significant concentra-tions of degradation products such as benzaldehydes, acetophenones, and aliphatic aldehydes.

Cao and Hewitt (1994) found that such build-up of degradation products is even worse for Car-botrap with the formation of benzene, and toluene, and to an even greater extent for Chromosorb 106 and thus it is not suitable for trapping volatile compounds at low concentrations. Nevertheless, Whitfield et al. (1983) used Chromosorb 105 for collecting TMA. Thomas and Rhue (1997) used Supelco ORBO-32 sorption tubes containing 0.65 g activated coconut charcoal for arsine trapping, but the analysis was not done conserving species by thermal desorption. It was done by solvent desorption with HNO3. Thus, despite the advantages of sorbent tubes concerning low costs, easy

Field sampling 53

handling, pre-concentration, and preservation of species information, no really successful applica-tions for volatile metals or metalloids are known.

The solid-phase micro-extraction (SPME) technique was first described by Arthur and Pawliszyn (1990). The commercial SPME design is similar to a GC injector syringe with a fused silica fiber of approximately 1 cm length coated with different stationary phases. Table 5 gives an overview of stationary phases available to date (07/2004), their properties, and application range. Liquid phase coatings, such as polydimethylsiloxane (PDMS) show sorption characteristics with extraction ca-pacity proportional to the coating volume. Solid coatings, such as Carboxen (CAR) or divinylben-zene (DVB), extract analytes by sorption with the extraction capacity proportional to the coating´s total surface area. The fiber´s cylindrical surface geometry is well defined and less complex than the surface in sorption tubes, facilitating more efficient extraction and desorption.

Table 5 Available SPME fibers and their properties (available date as of 07/2004; op. temp. = recommended operating temperature, cond. temp. = conditioning temperature, cond. time = conditioning time; non-bonded phases are stable with some water-miscible organic solvents, but slight swelling might occur, bonded-phases are stable with all organic solvents, slight swelling may occur when used with some non-polar solvents; PDMS = polydimethylsiloxane, DVB = divinylbenzene, CAR = Carboxen, PA = polyacrylate, CW = Carbowax, TPR = templated resin; Supelco catalogue information)

stat

iona

ry p

hase

film

thic

knes

s

desc

riptio

n

pH

max

. tem

p. [°

C]

op. t

emp.

[°C

]

cond

. tem

p. [°

C]

cond

. tim

e [h

]

appl

icat

ion

rang

e PDMS 100 µm

(GC/HPLC) Non-

bonded 2-10 280 200-280 250 0.5 low molecular weights (M = 60-275), volatiles

PDMS 30 µm (GC/HPLC)

Non-bonded 2-11 280 200-280 250 0.5 higher molecular weights (M =

80-500), non-polar, semivolatile

PDMS 7 µm (GC/HPLC) Bonded 2-11 340 220-320 320 1 high molecular weights,

non-polar M = 125-600

PDMS/ DVB

65 µm (GC) 60 µm (HPLC) Bonded 2-11 270 200-270 250 0.5

smaller molecules, volatile polar analytes (alcohols, amines, nitroaromates), M = 50-300

PA 85 µm (GC) Bonded 2-11 320 220-310 300 2 polar semivolatile analytes (M = 80-300) from polar samples

CAR/ PDMS

75 µm (GC) 85µm (GC) Bonded 2-11 320 250-310 300 1-2 smaller molecules, traces of

highly volatiles, M = 30-225 CW/ DVB

65 µm (GC) 70 µm (GC) Bonded 2-9 260 200-250 220 0.5 smaller molecules, alcohols,

polar compounds, M = 40-275 CW/ TPR 50 µm (HPLC) Bonded --- --- --- --- --- surface active substances

PDMS/ CAR/ DVB/

50/30 µm (GC) Bonded 2-11 270 230-270 270 1

smaller molecules, screening of a large range of molecular

masses (C3-C20), M = 40-275

Sorption tubes quantitatively sorb the analytes from the gas stream, but the SPME technique is based on the adjustment of a dynamic equilibrium between the concentration in the aqueous sam-ple, in the gaseous headspace, and on the fiber according to Fick´s second law (Zhang and Pawl-iszyn 1993):

54 Arsenic

2i

2i

ix

t)(x,cDt

t)(x,c

∂

∂=

∂∂

ci (x, t) analyte concentration at time t and position x in the aqueous sample (i = aq), in the gaseous headspace (i = g) or on the fiber (i = f)

Di diffusion coefficient in the gaseous headspace (i = g) and on the fiber (i = f; only diffu-sion along the fiber is considered, radial diffusion is assumed to be infinitely fast)

The two techniques “immersed SPME” (Arthur and Pawliszyn 1990) and “headspace SPME” (Zhang and Pawliszyn 1993) are used for sampling aqueous and gaseous environments, respec-tively. With the headspace SPME technique, equilibrium is generally reached more quickly (for As in 30-60 s on average) because diffusion rates are approximately 4 orders of magnitude higher in gas compared to water. Also, there are no matrix effects and the fiber is protected from swelling or degradation that might occur upon exposure to acid or highly mineralized solutions during im-mersed SPME.

Using a fiber with a liquid coating, the amount of analyte absorbed on the SPME fiber in the 3-phase system aqueous solution - gaseous headspace - fiber coating can be predicted from the fol-lowing equation (Zhang and Pawliszyn 1993):

aqggaqfgaqfg

fggaqaqf0aq

s VVKVKK

KKVVCn

+⋅+⋅⋅

⋅⋅⋅⋅=

ns amount of analyte absorbed by the fiber coating c0

aq initial concentration of analyte in aqueous sample solution Vf volume of fiber coating Vaq volume of aqueous sample solution Vg volume of gaseous headspace Kgaq partition coefficient of analyte between aqueous solution and headspace Kfg partition coefficient of analyte between headspace and fiber coating

Special narrow-bore liners (id = 0.8 mm, compared to 2-3 mm normally) are used for GC injection, guaranteeing a high laminar flow of the mobile phase along the fiber with sharp peaks and no re-peated sorption and desorption effects.

Mester et al. (2000) investigated the suitability of a 100 µm PDMS and a 75 µm PDMS/CAR fiber for the sorption of volatile As, Se, Sb, and Sn. They found that both higher extraction efficiency and faster desorption were achieved with PDMS/CAR compared to PDMS fibers. The slower de-sorption from PDMS is attributed to decomposition of the volatile compounds and consequent en-hanced interaction with the polymer coating. The same research group successfully applied PDMS/CAR fibers for sampling chloroarsines (Mester and Sturgeon 2001, 2002). According to experiments by Sur (1999), AsH3 shows only a low affinity to SPME fibers due to its low boiling point. Trimethylarsine could be successfully trapped on a PDMS/CAR/DVB fiber. However, fray-ing of this ternary fiber was encountered. Thus, the more stable PDMS/CAR fiber was chosen even

Field sampling 55

though signal intensities were only 78% of those on PDMS/CAR/DVB. The polar PA (Polyacrylat) and a bare PDMS fiber showed almost no trapping efficiency with 1 and 2%, respectively, of the signal intensity for PDMS/CAR/DVB.

The SPME technique has several advantages. Besides low costs, easy handling, and fast sample processing, it also has few interferences. With headspace SPME technique, matrix effects are eliminated and contamination potential and analyte losses are minimized. Using HG (section 2.6.1.1) with SPME trapping instead of continuous HG, interferences from excess hydrogen are eliminated because hydrogen is not efficiently trapped by the fiber at room temperature. One major disadvantage of the SPME technique is the fragile nature of the thin fibers, especially for applica-tions in the field. A further development beyond SPME application is the so-called stir bar sorptive extraction (SBSE) that consists of magnetic stir bars sealed in glass and coated with 50-300 µL PDMS (Baltussen et al. 1999). Compared to <0.5 µL PDMS normally used for SPME fibers, this method increases sampling capacity and, consequently, sensitivity 100 to 1000-fold. It might thus be very attractive for sampling volatile metallics of low concentrations, but no applications in this field have been found so far in literature.

2.7.4.3 Cryotrapping

Another method of gas trapping combined with analyte pre-concentration is cryotrapping. The gas samples are purged through a column packed with a porous polymer material (e.g., Chromosorb (10% SP-2100, 60/80 mesh) and cooled in liquid nitrogen or acetone to about -80°C (Feldmann et al. 1994, Feldmann and Kleinmann 1997). The temperature of -80°C, instead of the commonly used -196°C, prevents trapping of CH4 and CO2 which could block the cryotrap by condensation thus limiting sampling volume and negatively affecting the plasma stability when using ICP-MS. Alternatively, Feldmann et al. (2001) and Maillefer et al. (2003) used NaOH pellets for CO2, H2S, and H2O removal prior to trapping the target compounds on the Chromosorb column. Although this worked well for As, unstable or slightly acidic compounds such as SbH3 could react with NaOH. Cartridges filled with drying sorbents like CaSO4 (Maillefer et al. 2003), K2CO3, CaCl2 or Mg(ClO4)2 (Feldmann 1995) can also be used for H2O removal. After trapping, Feldmann et al. (1994), Feldmann and Kleinmann (1997), and Maillefer et al. (2003) closed the column with Swagelock stoppers, stored it in liquid nitrogen, and directly implemented it into the capillary col-umn of a GC.

The major advantage of cryotrapping is its very low selectivity, enabling the pre-concentration of highly volatile compounds. The disadvantages are possible matrix effects from other gas compo-nents such as CO2, CH4, H2S and possible losses due to the condensation of water and co-condensation of semivolatile compounds. Also the cryotrapping requires a costly setup, carrying a pump, liquid nitrogen, and a power supply to the field. The liquid nitrogen presents a possible haz-ard during transport. The vicinity of a laboratory is required for rapid analyses after sampling.

56 Arsenic

2.7.5 Techniques for collecting volatile metallics

In the previous sections several methods for trapping and stabilizing the volatile metallics with their advantages and disadvantages were presented. The initial task, however, is, to collect the vola-tile metallics in natural environments. Although sampling from the gas phase is relatively simple, it is much more complicated when volatile metallics dissolved in an aqueous phase are sampled and a gas-water separation has to be achieved.

2.7.5.1 Collection of volatile metallics from a gas phase

Well defined point sources such as gas wells on waste heaps, mine gas wells, and fumes in digester plants, can be sampled simply by connecting tubes to the gas outlet. Diffusive gas outlets can be sampled by different kinds of simple gas chambers. Feldmann and Cullen (1997) collected landfill gas which was visibly bubbling through a rain water puddle under a Plexiglas cylinder floating on a Styrofoam ring over the puddle. Hirner et al. (1998) used the same device for sampling geothermal gases bubbling through hot springs and pools. Maillefer et al. (2003) used an inverted plastic dish of 20 cm diameter, and 10 cm height for gas collection from underneath a compost pile. For sam-pling volatile Se from soils and plants, Biggar and Jayaweera (1993) used an acrylic volatilization chamber that was inserted approximately 10 cm into the soil and isolated a soil column with its plants. Similarly, Thomas and Rhue (1997) constructed a Plexiglas box with an internal air pump to trap surface soil gas flux of As from contaminated cattle dipping solutions.

Klusman (1991) developed and patented (US patent No. 4,993,874) an in-situ method for passive collection and immediate stabilization of volatile metals or metalloids. It was especially designed for As, Sb, and S, in soil gas or in the atmosphere above the ground as indicators for the explora-tion of subsurface mineral, geothermal and petroleum anomalies. Collectors with an oxidizing liq-uid inside are buried underneath a plastic shielding cone in the soil at a shallow depth. Apertures in the threaded plug in the top portion of the collector allow entry of ambient gases that are dissolved and oxidized in the trapping solution. Clean, high-quality glass wool is used to increase the surface area of the oxidizing liquid. After a certain period of time, the collectors are retrieved, sealed, and transported back to the laboratory. According to a written communication with Ron Klusman in June 2001, the efficiency of the apparatus was low based on laboratory calibration done with H2S to produce S(VI) in H2O2 as oxidizing liquid. The actual volume sampled is not known, since it depends on several parameters, like porosity, permeability, percentage of water filled porosity, gas solubility, oxygen exchange, and other parameters hard to determine. Results could only be inter-preted quantitatively by comparing a large number of collectors used in a survey with the same climate. Natural barometric gradients over sampling time were used as a “passive pump” for soil gas flow into the collectors. Applying a vacuum would have improved collection efficiency signifi-cantly but it entrained soil dust, contaminating the collector. One application for the patented method was an investigation on volatile As in soils at Ron Phibun, Thailand (Wongsanoon et al. 1997). Further work on this equipment was suspended.

Field sampling 57

2.7.5.2 Collection of volatile metallics dissolved in an aqueous phase

For sampling volatile metallics dissolved in an aqueous phase, little suitable equipment is available. A common technique for sampling gases dissolved in water is to take the water sample with a sub-mersible pump and analyze its gas content in the laboratory. Degassing prior to analysis has to be avoided by gas-tight sampling equipment and sample storage. In the laboratory the gas can either be extracted quantitatively from the water by vacuum application, or purging with an inert gas or obtained by setting up a dynamic equilibrium between the water sample and a gas phase above (headspace technique). The amount of analyte in the headspace depends on its Henry constant.

Amouroux et al. (1998) used a method, they called “in-situ purging” to sample volatile metallics (Se, Sn, Hg, Pb) dissolved in natural waters. The water sample is taken by a PTFE (Polytetrafluore-thylene) coated sampler that can be opened in a certain sampling depth, and transferred through a silicon tubing into a gas-tight 1 L pyrex bottle without headspace. Within 30 minutes after sample collection, the sample is purged through a moisture trap maintained at -20°C by a mixture of ace-tone and ice to remove water vapor, then finally condensed into the cryogenic trap immersed in liquid nitrogen at -196°C. Hirner et al. (1998) similarly took geothermal water samples in sealed teflon tubes, and purged them with helium in the laboratory for dissolved volatile compounds. These purging methods, however, have some serious disadvantages. First, they are non-integrative determinations with the drawback of short-term temporal noise as described above. To achieve low detection limits, relatively large amounts of water have to be taken. Exact depth-dependent sam-pling is difficult and presents a major disturbance for the sampled water body. The sensitive gas equilibrium might further be disturbed during storage and while purging the sample. Especially for studies focusing on the diffusive degassing at the sediment - water interface in-situ methods are required. However, such methods are still rare.

So-called passive diffusion samplers eliminate the problem of non-integrative sampling and possi-ble disturbances of the gas equilibrium. After installation they are left several days to weeks outside while the air in the sampler equilibrates with the gaseous compounds in the surrounding aqueous phase. Church et al. (2002) developed passive vapor-diffusion samplers (empty glass vials enclosed in two layers of polyethylene membrane tubing) to detect volatile organic compounds in discharge areas of contaminated groundwater. However, the passive diffusion samplers include no pre-concentration or on-site stabilization and are hardly suitable for the generally low concentrations of volatile metallics. Furthermore, equilibration times of up to 3 weeks (Church et al. 2002) are not practical for fast screening. According to Klusman (1991), the patented method described above for in-situ collection and stabilization of volatile metallics can also be used for gas collection from the sediment-water interface at the bottom of a shallow water body. The shielding cone is then weighted to hold it to the bottom and used like a diving bell. A nylon mesh screen that is secured to the covering cone supports the collector. For deeper water application, the author recommends to add air or inert gas as the assembly is lowered to compensate for compression during descent. The whole application, however, is rather impractical with a high potential for flooding the collector. No applications of this method were found in literature.

58 Arsenic

2.8 Summary of chapter 2

Section 2.5 has given a detailed overview over some of the most predominant As species in aquatic environments, illuminating their significant differences in occurrence, mobility, stability and toxic-ity, as well as the potential of chemical and biologically mediated species conversions. As early as in the mid of the 19th century, focus has already been drawn to accidental releases of volatile As (section 2.5.3). Species-selective determination methods, however, have not been introduced until the early 1970s with the invention of the hydride generation technique (section 2.6.1.1). Within the last 10 years, considerable effort has been paid to scrutinizing the As biomethylation pathways in different organisms (section 2.5.2). However, comprehensive studies on the occurrence of dis-solved methylated and especially volatile As species in natural aquatic environments are still rare. Quantification has seldom been achieved, and several, especially volatile compounds, like sulfur (section 2.5.2.5) or chloroarsines (section 2.5.2.6), have not been even qualitatively identified in natural systems, yet.

The problem with identifying volatile As compounds in the environment is partly attributed to a lack in standardized sampling procedures. A wide variety of tests and methods was presented (sec-tion 2.6), however most of them are not available commercially as standard equipment, but custom-made coupled methods that are specifically developed by different research groups mostly from the field of analytical chemistry. Different opinions exist about the stability of dissolved and volatile As species, and various on-site trapping methods have been proposed (section 2.7). Some of them, however, are only applicable if analysis can be guaranteed within a few hours of sampling. Few suggestions exist for sampling remote areas. A fast and efficient method for in-situ sampling of volatile As species from the aqueous phase is still lacking (section 2.7.5.2).

The present thesis therefore focused on the development of a robust technique for sampling volatile metallics in aquatic systems (section 3) as well as on investigating from a hydrogeochemical point of view geothermal features at Yellowstone National Park, as an example for an aqueous environ-ment predestined for the release of volatile metallics (section 4).

3 DEVELOPMENT OF A FIELD METHOD

To sample volatile metallics from diffusive sources in aquatic environments two different methods were applied; field practicability was the major criterion. For a first screening on total concentra-tions of volatile metallics, a method was developed with an in-situ gas-water separation using a gas-collector cell with a porous PTFE (Polytetrafluorethylene) membrane (section 3.1), trapping the volatile metallics in a liquid sorbent (background section 2.7.4.1; Planer-Friedrich et al. 2002). The advantage of this screening method is its low cost, easy handling, applicability in the field, efficiency in separating gas samples from aqueous environments, long-time stability of the dis-solved gas samples, and the possibility for quantitative interpretation (section 3.2.2). The disadvan-tage is that species information is lost due to hydrolysis and oxidation of the volatile metallics. Therefore, the suitability of sorption tubes and SPME technique (background section 2.7.4.2) fol-lowed by GC-MS analysis was investigated in the laboratory to be used in areas where significant total concentrations of volatile metallics were found during screening (section 3.2.3).

3.1 Development of an in-situ gas-collector cell

For efficient gas-water separation, a gas-collector cell was designed that can be placed inside a water body or at the sediment-water interface. It is connected to the trapping unit with liquid or solid sorbents to be described in section 3.2. The material used for the gas-collector cell is PTFE (Polytetrafluorethylene, [-F2C-CF2-]n). It was discovered by Roy Plunkett at DuPont in 1938 and is commonly known by the DuPont brand name Teflon®. Besides its chemical inertness and low ad-hesion, it is one of the most hydrophobic materials with a surface tension of 18 dynes/cm, com-pared to 73 dynes/cm for water making it ideal for efficient gas-water separation.

Three different prototypes were designed and constructed in a workshop of the Technische Univer-sität Bergakademie Freiberg ( Figure 15 - Figure 17). For the first prototype ( Figure 15 right, and Figure 16 right) the collector cell was drilled from teflon bars with different porosities (5 µm and 10 µm, porous PTFE from Bohlender). Two different side thicknesses, 3 mm and 6 mm, were cho-sen. The cells are closed by a non-porous teflon cap with a screw connection for a 2 mm teflon capillary leading collected gases from inside the gas-collector cell to the trapping unit. The cap is attached to the porous cell below by two stainless steel rings, kept together by 4 screws. A greased O-ring inside seals the connection gas-tight. The second prototype ( Figure 15 left, and Figure 16 left) consists of a porous teflon tube of 10 µm porosity, 6.4 mm thickness, an outer diameter of 52.4 mm, and an inner diameter of 39.6 mm (PTFE from Reichelt Chemietechnik). On top and bottom the tube fits in small joints of non-porous teflon plates (PTFE from Reichelt Chemietech-nik). The plate on top contains a screw connection for a 2 mm teflon capillary to lead the collected gases to the trapping unit. Three 6.5 cm long titanium screws keep the upper and lower lid together. Threads are drilled in the titanium plates supporting the teflon lids.

The gas-collector cells were kept small (5 and 6 cm high, 3 and 5 cm wide) to allow for an accurate depth-orientated sampling. Inserting these gas-collector cells in the aqueous medium, only gas dif-

60 Development of a field method

fusion takes place through the porous teflon surfaces. Water does not penetrate due to the hydro-phobic properties of teflon. Applying vacuum to the gas-collector cells increases sampling effi-ciency. At a vacuum exceeding -50 mbar, even though all connections were still waterproof, the bubbling point of the porous teflon was reached and water breakthrough occurred. The different designs with porosities of 5 and 10 µm, and wall thicknesses of 3 and 6 mm did not show signifi-cant differences in water breakthrough points.

Figure 15 Gas-collector cells prototype 1 (right) and 2 (left), for collecting gases from diffusive sources in aqueous environments applying no or low (-50 mbar) vac-uum; teflon capillaries lead the col-lected gases from inside the gas-collector cell to the trapping unit

Figure 16 Gas-collector cells prototype 1 (right) and 2 (left); mate-rial: porous PTFE with a porosity of 5 and 10 µm and a side thickness of the 3 and 6 mm

A third prototype was developed to apply higher vacuum for increased sampling efficiency ( Figure 17). In contrast to the first two designs, gas transport does not take place through pores over the entire body of a hollow porous teflon tube, but through a teflon membrane (PTFE from Infiltec, porosity 0.1 µm), placed in a filter of solid teflon. The outer side is protected towards the sampling medium by a porous teflon disk (PTFE from Bohlender, porosity 10 µm, thickness 5 mm) and to-wards the inner side by a solid teflon disk (PTFE from Bohlender, thickness 6 mm) to withstand vacuum-induced deformation. To allow gas passage, the solid teflon disk contains numerous 1 mm holes, similar to a sieve. This gas-collector cell withstands vacuum up to -500 mbar (Infiltec even cites applications with > 2 bar pressure) without breakthrough of water. Figure 18 shows a setup of the described equipment including gas-collector cell, liquid sorbent and vacuum container. Vacuum

Gas generation and trapping 61

is applied only once and the sampling set is left out in the field until the vacuum is completely diminished. Concentrations of the volatile metallics in the gas phase [ng/m3] can then be calculated from the mass of volatile metallics trapped in the liquid sorbent [µg] and the replaced vacuum vol-ume [m3]. Connecting a battery-powered pump would be an option for continuous low flow gas sampling.

Figure 17 Gas-collector cells proto-type 3 for collecting gases from diffusive sources in aqueous environments applying vacuum up to -500 mbar; material: PTFE membrane (M) with a porosity of 1 µm, placed in a filter of solid PTFE, protected to the sampling media below by a porous PTFE disk (porosity 10 µm, thickness 5 mm, left of the membrane) and above by a solid PTFE disk with numerous 1 mm holes for gas transport to withstand de-formation because of vacuum application (to the right of the membrane)

Figure 18 Field setup of the sampling and trapping equipment showing the gas-collector cell (CC) installed close to the surface of the water saturated sediment, the trapping and stabilizing unit (liquid sorbent, LS), and a vacuum container (VC)

3.2 Gas generation and trapping

Volatile metallics were produced by HG (section 2.6.1.1) in the laboratory, to test the efficiency of the trapping unit connected behind the gas-collector cell. Figure 19 shows the laboratory setup. A high-grade stainless-steel reaction vessel (250 mm high, 100 mm diameter) was constructed as an abiotic reactor symbolizing the natural aqueous environment, where volatile metallics are pro-duced. PTFE coating of the steel proved to be necessary to avoid sorption effects and precipitation

Valve connection for vacuum pump

VC

CC

LS

M


because of pH-changes during HG. The reaction vessel is filled to about half its height with sample solution (475 mL). Chemical reagents for abiotic hydride generation are added with a syringe and a capillary connected to the bottom of the reaction vessel.

A porous ceramic disk (6 cm diameter, 0.6 cm height) was initially used on the inside bottom of the reaction vessel for finer distribution of the added chemical reagents. Cleaning the disk in 1 N HCl after 24 experiments with a total supply of 18.3 mg As, yielded a total of 0.75 mg As, equivalent to an accumulative sorption of 4%. Because hydride generation starts at a pH of about 1.2, some of the As sorbed on the ceramic disk is probably released at the beginning of each new experiment and precipitated again with increasing pH. Complete desorption during each experiment on the other hand can be excluded because the last experiment before cleaning the ceramic disk yielded 0.073 mg total As only compared to 0.75 mg detected in the cleaning solution. More likely, it is a dynamic equilibrium, where the 0.75 mg As sorbed may already present the maximum sorption capacity of the ceramic disk. However, to be on the safe side, the disc was dismissed in the follow-ing experiments and the reagents were added to the reaction vessel directly through a hole in the bottom. The gas-collector cell is placed inside the reaction vessel. Gases created in the reaction vessel pass the teflon membrane of the gas-collector cell and finally leave the vessel to the trapping unit by a teflon capillary.

Figure 19 Laboratory setup for generation of volatile metallics in a PTFE coated reaction vessel (RV), consequent sampling with the devel-oped collector cell (CC) and transfer to the liquid sorbent (LS); syringe (SY) is used to add reducing agent and acid for hydride generation, N2 pro-vides reducing environment during the whole experiment, pressure is moni-tored by a digital manometer (MA).

3.2.1 Optimization of arsine generation

Dimethylated arsines created from dimethyl arsenic sodium salt trihydrate (C2H6AsNaO2⋅3H2O; FLUKA p.a. ≥ 98%) were used to optimize the method. The optimized conditions found applica-tion for additional experiments with inorganic arsine AsH3 (created from AsVHNa2O4⋅7H2O, FLUKA, p.a. ≥ 98%) and proved the method´s efficiency. Trapping other volatile metallics should perform in the same way. Because optimum generation of other volatile metallics requires addi-

RV

CC LS1 LS2

MA SY

N2


tional time-consuming adjustments, explicit confirmation of these conditions was dispensed. NaBH4 (p.a. ≥ 96%, MERCK) in NaOH was used as a reducing agent.

Small-scale derivatization experiments yielded the best relative response for arsine generation from DMAA in a pH range between 0.4 and 1.5, with an optimum at pH 1.3 ( Figure 20) in general ac-cordance with other literature discussed in section 2.6.1.1. Even though HG is a standard method, common volumes are significantly less than 1 mL. For the experiments presented here, the goal was to test sampling efficiency in an aqueous environment using the developed collector cell, re-quiring water volumes of approximately 0.5 L in the reaction vessel. Problems were encountered with simply upscaling the amount of chemicals needed for HG. Few reports were found in litera-ture dealing with HG from large sample volumes. Luna et al. (2000) reported no signals for Ag, Cu, Au and Zn in hydride generation when relatively large volume (> 1 mL) batch reactions were attempted. The authors presented no reasons or solutions for this drawback.

Figure 20 Experiments for DMA HG from 10 mL deionized water containing 20 µg/L As as C2H6AsNaO2⋅3H2O. Reduc-tion was performed for pH of 0.05 to 5.5 by adding variable amounts (0 to 2.5 mL) of 30% HCl and 2.4 mL 3% NaBH4 in 1% NaOH

To check the efficiency of hydride generation in the reaction vessel under various conditions, sam-ples were taken from the original standard solution prior to the experiment (samples “C” = control) and from the solution in the reaction vessel after the experiment (samples “R” = remaining). The percentage of volatilized As was calculated according to the following equation:

ed volatiliz% 100VcVc% 100

CC

RR =⋅⋅⋅

−

c = concentration V = volume of initial solution in reaction vessel (C), and solution after reaction in reaction vessel (R)

All samples were analyzed for total As by GF-AAS (graphite furnace AAS EA4, Zeiss, Germany) using platform technique and the conditions given in Table 6, section 3.2.2.

Using FI-HG, a constant purging gas stream transfers the created volatile compounds directly to the detection unit. The PTFE membrane of the collector cell slows down the gaseous flow, leading to longer residence times of volatile compounds in the reaction vessel. Because the reducing agent NaBH4 is prepared in NaOH for stability reasons, the pH increases significantly during the creation of the volatile compounds. Not all As is completely reduced to dimethylarsine and degassed from

0102030405060708090

100

0 1 2 3 4 5 6pH

rela

tive

resp

onse

DM

A [%

]


solution. The remainder is re-transformed to dissolved Na(CH3)2AsVOO once the solution is alka-line again.

During the first experiments ( Appendix 2: experiments No.1-23), a vacuum pump was used for gas flow control. Because HG increases the pressure inside the reaction vessel, the injection times for the reducing agent and the experiment´s total time were adjusted to keep the differential pressure at about 60 mbar, resulting in extremely variable HG rates. The shorter the total reaction times, the lower was the percentage of As volatilized. The experiments 2, 4, and 8 ( Appendix 2) yielded vola-tilization rates of 45, 41, and 63% only with total reaction times of 320, 185, and 435 min, respec-tively, compared to volatilization rates of 70-76% for total reaction times > 500 min ( Appendix 2: 1, 3, 5-7, 9). Increasing NaBH4 concentrations, and decreasing HCl concentrations ( Appendix 2: 10-23) increased average volatilization rates to 72-89%; the highest NaBH4 concentrations yielded 95-96% As volatilization ( Appendix 2: 11, 12), the lowest HCl concentration 98% ( Appendix 2: 18).

Because oxidation of the generated arsines was thought to be the major cause for low volatilization rates, a continuous N2 flow was applied to maintain reducing conditions during the further experi-ments ( Appendix 2: 24-105). After initial experiments with N2 pressures of 800 to 200 mbar, N2 pressure was adjusted as low as possible (approximately 30 mbar differential pressure) to keep the pressure inside the reaction vessel below the gas transfer rate limited by the permeability of the PTFE membrane. The N2 flow was interrupted only during the few seconds of reactant injection.

Finally, a two-step strategy of reduction, re-acidification and second reduction proved to yield complete and reproducible volatilization ( Figure 21; Appendix 2: 29, 41-105). The primarily acidi-fied DMAA standard solution with a pH of approximately 1.2 in the reaction vessel was first purged with N2 for 10 minutes to create reducing conditions. Flow from the bottom to the top al-lowed for complete oxygen removal. The reducing agent NaBH4 was used in a concentration of 15% in 5% NaOH. As Shraim et al. (1999) already detected, concentrations >1% NaBH4 result in turbid solutions. The non-dissolved particles settle down and the solution becomes clear with time. However, when mixed with acid solution during HG, the particles re-dissolve producing a non-homogeneous reaction mixture with localized cells of NaBH4 concentrations. To reduce this insta-bility, the solution was filtered before usage.

The reducing agent was added discontinuously over 20 minutes in steps of approximately 2 mL every 2 minutes. This stepwise addition of small amounts of reducing agent was chosen because an intense arsine generation at the beginning significantly increased the pressure in the vessel due to the limiting permeability of the PTFE membrane, partly leading to backflow through the bottom capillary. The reaction vessel was purged with N2 for another 5 minutes after the first reduction step; pH was at about 12. The solution was acidified again (pH 1 to 2) by adding 40 mL of 1 N HCl over 20 minutes. All As, not yet degassed and re-dissolved in solution, was transferred again to (CH3)2AsO(OH) and reduced in a second step with 20 mL 15% NaBH4 in 5% NaOH inserted over


15 minutes. A final N2 flow for 20 minutes ensured that all volatile compounds were volatilized from the reaction vessel via the PTFE membrane. The whole experiment took 90 minutes.

Figure 21 Two-step strategy of reduction, re-acidification and second reduction for complete and repro-ducible HG of volatile arsines from 475 mL of a dimethyl arsenic sodium salt trihydrate solution at a continuous N2 flow of 30 mbar; single bars indicate stepwise injection of reactants (HCl, NaBH4)

With the described two-step approach of two times acidification and reduction, As concentrations after reaction were close to or below the detection limit of 1.5 µg As/L. Volatilization rates gener-ally were >95%. Further experiments to decrease the amount of chemical reagents ( Appendix 2 E 31-33) or shorten the reaction time ( Appendix 2 E 34-37) led to incomplete transfer from the aque-ous to the gaseous phase (85-90% volatilization). Not acidifying the solution initially ( Appendix 2 E 84) yielded volatilization rates as low as 41%. No difference was found in volatilization effi-ciency between the collector cell prototypes 1 and 2 with gas flow through porous (5 and 10 µm) Teflon bodies ( Appendix 2 E 48-51) and prototype 3 with gas flow through the Teflon membrane (porosity 0.1 µm).

3.2.2 Optimization of screening on total concentrations (liquid sorbents)

Four different reaction vessels, and three different oxidizing agents were tested for their suitability of trapping total volatile As. The amounts of liquid sorbent were kept small to achieve an optimum enrichment of the naturally low concentrations of volatile metallics. The reaction vessels were col-umns of small diameter and at least 12 cm height to guarantee sufficient reaction time while the gas is bubbling through the oxidizing solution. Small glass and teflon bodies, so-called Raschig rings, were used to further increase reaction surface and reaction time. Trapping efficiency in liquid sorbent is calculated according to the following equation:

fraction ed volatiliz theof trapped% 100VcVc

Vc

RRCC

OO =⋅⋅−⋅

⋅

c = concentration V = volume of initial solution in reaction vessel (C), and solution after reaction in reaction vessel (R),

and oxidation solution (O)

0

5

10

15

20

25

30

35

40

0 10 20 30 40 50 60 70 80 90experimental time [min]

0

2

4

6

8

10

12

14

pH

reac

tant

vol

ume

[mL]

N2 N2 N2

pH pH

1st reduction 2nd reductionre-acidification

1 N HCl

NaBH4 in NaOH NaBH4 in NaOH


Laboratory experiments began with a commercial glass wash bottle (250 mL volume, 14 cm height) as oxidizing trap. However, the lid did not prove to be completely gas-tight under the high pressure created during hydride generation. Volatile As was observed by garlic-like odor. The oxi-dizing agent, hydrogen peroxide (H2O2, perhydrole p.a. 30% MERCK), was not suitable for rapid oxidation, yielding trapping rates of 3-17% only ( Appendix 3 E 1-23). No significant improvement was found for different H2O2 strengths (5% to 30% in the final solution) or making basic the oxi-dizing solution with 0.1 and 1 N NaOH. The low trapping efficiency was even worse than that re-ported by Wongsanoon et al. (1997) who deducted trapping efficiency for volatile metallics in 30% H2O2 from tests with hydrogen sulfide to be about 50%.

Using a stainless steel cylinder (40 cm high, 4.5 cm diameter) as an alternative oxidizing trap, in-stead of the glass wash bottle, H2O2 showed slightly higher trapping rates of 10, and 22% ( Appendix 3: 24, 25). A 1.5 M sodiumperoxodisulfate (Na2S2O8, p.a. ≥ 99% FLUKA) solution as oxidizing agent ( Appendix 3: 26) caused problems due to high matrix effects during the following analysis by GF-AAS. Further experiments with Na2S2O8 were dismissed. Sodium hypochlorite (NaOCl, 6-14% Cl active, Riedel-de-Haen) turned out to be effective for rapid and complete oxida-tion. However, a special temperature-time-program had to be developed to analyze the complex matrix of the NaOCl-samples by GF-AAS ( Table 6).

Table 6 AAS conditions for analyzing the original standard solution before the experiment (samples “C” = con-trol), the solution in the reaction vessel (samples “R” = remaining) and the oxidizing solution (samples “O”)

STEP TEMP [°C] RAMP [°C/s] HOLD [s] TIME [s] RUNDrying 120 10 0 10.1 Drying 150 3 30 40.0

Dry ashing 1350 C 1400 R 1500 O

135 C 140 R 150 O

30 C 40 R 60 O

38.9 C 48.9 R 69.0 O

Atomization 2400 FP (full power) 4 4.7 C 0.2 R0.2 O

Cleaning 2600 1000 3 3.2 Note: 15 µL sample solution + 8 µL modifier (Pd(NO3)2; MERCK c(Pd) = 10 g/L Pd(NO3)2 in ca. 15% HNO3; dilution 1:10); wave length 193.7 mm, gap width 60 mm, standards for C-samples (initial solution before HG) in 0.1N HCl ma-trix, standards for R-samples (solution after HG) in 8 mL 1 N HCl and 19 mL 1 N NaOH matrix to 50 mL, standards for O-samples (oxidizing solution) in 11.1mL NaOCl and 11.1 mL HCl matrix to 50 mL

Non-specific sorption peaks appeared during analysis because of the high salt concentrations in these samples. The method development aimed at a complete separation of those peaks from the As signal by varying dry ashing conditions. High dry ashing temperatures and a total duration of about 70 s were used for this step. To avoid As losses during this thermal treatment an increased modifier concentration, 8 µL of a 1 g Pd/L stock solution (Pd(NO3)2), was added to 15 µL sample solution. Experiments with NH4NO3 to improve the separation of non-specific and specific signals showed no success. Furthermore 1 mL 1N HCl and 2.5 mL H2O were added to 1 mL of oxidizing sample ("O"-sample) due to problems with the high alkalinity and the applied high As concentrations.


Overall trapping efficiencies for NaOCl were between 53% and 86% in the stainless steel cylinder ( Appendix 3 E 27-36, 38). Acidifying pure NaOCl from pH 11.8 to pH 4.0 (1:1 NaOCl : 1N HCl; Appendix 3: 37) brought trapping efficiency down to 1%, probably due to degassing of reactive Cl during acidification of the oxidizing agent (strong Cl odor). Strongly alkaline oxidizing solutions (pH 12.3, 1:1 NaOCl : 1N NaOH; Appendix 3: 39) also yielded a lower than average trapping effi-ciency with 57%. The 90 min two-step approach optimized for arsine generation also proved the best results for trapping (79, 86%; Appendix 3: 38, 29), whereas single-step reduction yielded al-ways lower trapping rates (53-67%; Appendix 3: 27, 28, 30, 31). This result suggests that during the initial reduction and arsine generation combined with high pressure and rapid gas flow, the oxidation capacity of the trapping solution was exceeded. Stepwise injection and longer reaction times gradually transferred the arsines, leading to a better sampling capacity. This problem is mainly a laboratory artifact. During gradual biotic generation of volatile metallics in nature, sorp-tion capacities will not be exceeded in that way.

Glass beads in the NaOCl solution increased sampling efficiency from 53% ( Appendix 3: 30) to 58% ( Appendix 3: 28), and from 75% ( Appendix 3: 32) to 86% ( Appendix 3: 29). However, glass beads showed about 2% accumulative As sorption (12 µg from about 530 µg total As from 14 ex-periments). A control experiment with an As-free solution yielded 28 µg/L As in the stainless steel cylinder oxidizing trap filled with glass beads ( Appendix 3: 40). This blank value proved desorp-tion both from the stainless steel cylinder and the glass beads.

Therefore, a commercial teflon wash bottle (250 mL, height 13.5 cm, teflon = perfluoroalkoxy PFA from BOLA) replaced the stainless steel cylinder, and PTFE rings replaced the glass beads. The PFA bottle did not show sorption effects but the relatively soft material was not really suitable for field application nor for withstanding the built-up of high pressures during hydride generation in the laboratory. Volatile As losses at connections and threads were detected by garlic-like odor. Trapping rates were less than with the stainless steel cylinder (23-62%; Appendix 3: 41, 43-55). PTFE rings used instead of the glass beads showed an interesting phenomenon. The teflon rings were cut from teflon tubes with 6 mm or 2 mm inner diameter and 8 mm or 4 mm outer diameter (T6x8, respectively T2x4) and about 2-3 mm thickness. Using only T6x8 significantly increased trapping efficiency (77%; Appendix 3: 42) compared to experiments without Raschig rings (42, 45%; Appendix 3: 45, 41). When T2x4 were used, however, trapping efficiencies were low ( Appendix 3: 43, 44, 46-55). This result could indicate that inside the smaller teflon rings isolated gas bubbles form which can not interact with the oxidizing medium and also reduce further gas transport through the rings, thus decreasing reaction time inside the oxidizing solution. For the T6x8 rings the inner cavities are large enough to allow unimpeded gas flow. For all subsequent experiments, rings with an inner diameter of 4 mm (T4x6) were found to be a good compromise between optimum surface reaction and unimpeded gas flow.

As final oxidizing vessels, sturdy and inert PTFE bottles (20 cm high, 4 cm diameter) were de-signed and constructed in a workshop at the Technische Universität Bergakademie Freiberg. With a


teflon capillary, collected gases are conducted through the top lid directly inside to the bottom of the bottle without any further connections, thus excluding losses.

Different dilutions of the NaOCl solution were tested from undiluted to 1:500 NaOCl : deionized H2O, and a pure deionized water control solution ( Figure 22). Dilution ratios of 1 : 100 (0.06-0.14% active Cl) proved sufficient to obtain overall recovery rates of 84% ± 6.6 (86% ± 7.8 for HG from DMA only ( Appendix 3: 64, 71, 77-83), and 82% ± 2.9 for HG from As(V) only ( Appendix 3: 82 a-f). This result was confirmed for a concentration range from 6-150 µg As/L in solution be-fore HG ( Figure 23).

Figure 22 As recovery after volatilization from DMAA and As(V) solutions and trapping in NaOCl in different dilution ratios from 1:0 (pure NaOCl) to 0:1 (de-ionized water, control) (bars indi-cate minimum and maximum val-ues, diamonds indicate mean val-ues)

Concentrations in a second oxidizing solution (LS 2; Figure 19), mounted behind the first one (LS 1), were close to or below detection limit. Lower recovery rates for the more concentrated oxidiz-ing solutions (1:0, 1:1, 1:10), as opposed to the 1:100 solutions, are probably not related to insuffi-cient trapping but to problems with GF-AAS analysis because of high matrix effects. Dilutions of 1:200 and 1:500 led to incomplete trapping, the second oxidizing solution (pure NaOCl) mounted behind the first one showed higher As concentrations (up to 25%). A blind test with deionized wa-ter as LS1 showed As concentrations of less than 1% in LS1 and 86% in LS2 (pure NaOCl) ( Appendix 3: 67).

Dilution 1:0 68 % ± 12.2 ( Appendix 3: 56, 58, 59, 60, 61, 65, 68) Dilution 1:1 71 % ( Appendix 3: 69) Dilution 1:10 70 % ± 8.2 ( Appendix 3: 62, 70) Dilution 1:50 84 % ± 5.5 ( Appendix 3: 63, 66, 74, 76) Dilution 1:100 84 % ± 6.6 ( Appendix 3: 64, 71, 77-83, 82a-f) Dilution 1:200 45 % ( Appendix 3: 73) Dilution 1:500 43 % ± 5.1 ( Appendix 3: 74, 75) Dilution 0:1 <1% ( Appendix 3: 67)

0102030405060708090

100

1:0 1:1 1:10 1:50 1:100 1:200 1:500 0:1

NaO Cl : H2O dilution ratio

As r

ecov

ery

[%]


NaOCl : H2O ratio

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100 120 140 160

As concentration standard solution [µg/L]

As

reco

very

[%]

1:01:11:101:501:1001:2001:500

Figure 23 Volatile As recoveries from solutions of different DMAA and As(V) concentrations; symbols indi-cate varying dilutions of the NaOCl oxidizing solution ( Appendix 3: 56, 58-66, 68-83, 82a-f)

The long-term stability of the 1:100 diluted solutions is shown in Table 7. Maximum losses during 6 weeks storage in a refrigerator were 12% in glass bottles and 13% in PE bottles compared to re-sults obtained 1-2 days after the experiment. Mean losses were 5% in glass, and 8% in PE bottles. Even though the losses in glass bottles were slightly lower than in PE bottles, PE bottles are rec-ommended for further analysis in order to avoid sorption effects of other metallics on the glass walls, due to the high pH (pH = 11.1 - 11.3).

Table 7 Long-term stability of the 1:100 diluted solution comparing the analysis 1-2 days after the experiment to the analysis after 6 weeks of storage in a refrigerator; mean losses were 5% in glass and 8% in PE bottles

As concentrations [µg/L] As losses after 6 weeks [%] in glass bottles 1-2 days after the ex-

periment

in PE bottles 1-2 days after the ex-

periment

in glass bottles 6 weeks after experi-

ment

in PE bottles 6 weeks after ex-

periment in glass bottles in PE bottles 284 278 251 255 12 8 136 132 135 129 0 2 373 373 382 394 -3 -6 22 21 20 19 8 10 70 72 69 63 1 13


3.2.3 Species-selective sampling (solid sorbents)

The gas-collecting procedure for species-selective determinations is the same as described in sec-tion 3.1. The laboratory setup shown in Figure 19 with the optimized HG conditions discussed in section 3.2.1 was applied. The liquid sorbent traps (LS1, LS2 in Figure 19) are replaced by solid sorbents (sorption tubes, SPME fibers) as described in section 2.7.4.2. In some experiments, a liq-uid sorbent was mounted behind the solid sorbent to estimate the amount of volatile As break-through.

3.2.3.1 Sorption tubes

Sorption tubes were custom made from about 9 cm long glass tubes. The tubes´ outer diameter was set to 8 mm by the dimensions of the desorption port of the GC-MS used (Bruker mobile GC-MS EM 640). The inner diameter was 5 mm. Before packing the desorption tubes, they were acid cleaned, rinsed with deionized water and heated at 100°C overnight. About 26 mg of silanized glass wool (Glass Wool Silane Treated; Supelco) were inserted as a plug in the lower end of the tube that was tapered before. Five different types of sorption material were used: Tenax TA (60/80 and 80/100 mesh; SKC), Chromosorb 105 and Chromosorb 106 (60/80 mesh; Varian), Carbotrap (ORBO 100, 20/40 mesh; Sigma-Aldrich), Carbosieve S III (60/80 mesh; Scientific Instrument Services), and Carboxen 564 (ORBO 92, 20/45 mesh; Sigma-Aldrich) ( Table 4). A total amount of 150 to 250 mg sorption material resulted in packing heights between 1.5 cm and 3.5 cm. Larger bed sizes were avoided because they will trap more water and contribute more background noise to the chromatogram.

Single-bed resins ( Appendix 4 No.17-20), double-bed resins ( Appendix 4 No.1-9, 14-16) and tri-ple-bed resins ( Appendix 4 No.10-13) were designed. Multi-bed resins have the advantage that they can trap a wide variety of volatile compounds. They are packed so that the material that sorbs the smallest and most volatile analytes, forms the top layer of the multi-bed resin. Larger molecules are trapped on the bottom and do not block or irreversibly sorb on the finer sorption material. The direction of gas flow during desorption is reverse to that of sorption during sampling, guaranteeing that larger, less volatile molecules never pass through the finer material on top of the resin. Differ-ent packing materials in multi-bed resins were separated by 13 mg silanized glass wool. About 15 mg silanized glass wool were used as plug for the upper end of the resin together with a steel spring to keep the sorption material in place. Two tubes were filled with silanized glass wool only to test blind values in the absence of sorption material ( Appendix 4 No. 28, 29).

The upper end of the glass tube was tapered after packing and the desorption tubes were condi-tioned at 50°C over their final operating temperature for 4 hours in a continuous N2 flow (N2 6.0 = N2 >99.9999 Vol%). So as not to load the chromatographic column with contaminants from this cleaning step, a desorption apparatus with continuous gas flow was designed and constructed by a workshop at the Technische Universität Bergakademie Freiberg ( Figure 24). One GC run after the


conditioning step confirmed low blank levels. Besides the homemade sorption tubes, some com-mercial ones from SKC were selected ( Appendix 4 No.21-27, 30-32) and conditioned the same way.

Figure 24 Desorption apparatus for conditioning and cleaning sorption tubes; continuous gas flow (N2) can be applied both to the heat-ing unit (1) that provides space for desorbing 1 sorp-tion tube, and the cooling unit (2) with space for 10 sorption tubes to be cooled down under oxygen exclu-sion after heat desorption

Three different columns were used in the Bruker mobile GC-MS EM 640:

• DB-624 (6% cyanopropylphenyl-94% dimethylpolysiloxane) with a medium polarity, com-monly used for volatile organics, column length (L) = 30 m, inner diameter (ID) = 0.25 mm, film thickness (fd) =1.4 µm

• DB-5 (5% phenyl-95% dimethylpolysiloxane) non-polar, L = 12 m, ID = 0.32 mm, fd = 1 µm • MDN-12 (exact composition patented, no information) low polarity, L = 30 m, ID = 0.25 mm,

fd = 0.25 µm

With the DB-624 column no volatile arsine compounds could be detected in the chromatograms using sorption tubes with Tenax, Carbosieve, Carboxen, and Carbotrap beds ( Appendix 5: 85-105). Oxidizing traps were mounted behind the sorption tubes to determine the breakthrough of arsine through the sorption tubes. A significant As breakthrough was found for the commercial Tenax TA tubes, single-bed resins of Carbosieve, and Carbotrap, as well as for the Tenax-Carbotrap double-bed resin. From the original As in solution 84% ( Appendix 6: 85), 41% and 36% ( Appendix 6: 92, 100), 53% ( Appendix 6: 93), and 82% ( Appendix 6: 89) passed the sorption tube and were trapped in the oxidizing solution. Increasing the total sorption surface by mounting two Tenax TA resins behind each other did not improve sorption capacities. Arsenic recovery rates in the oxidizing solu-tion behind the sorption tube remained high with 82, and 92% respectively ( Appendix 6: 104, 105). For all the other experiments with double and triple-bed resins no As breakthrough was detected in the oxidizing solutions. To test the response without sorption material, two tubes were filled with silanized glass wool only. The results yielded 88% As recovery in the oxidizing solution behind the sorption tubes ( Appendix 6: 103). After the experiment, the sorption tubes were destroyed and the

1

1 sorption tube - heating

2 ports occupied by sorption tubes - cooling

N2

excess heat

thermometer2

lid for heating unit


glass wool was eluted with 1 N HCl. Arsenic concentrations below detection limits proved that the silanized glass wool was inert and no arsine was trapped on it.

Both the non-polar MDN-12 and the DB-5 columns proved to be suitable for arsine separation. The optimized GC-MS conditions are shown in Table 8. Desorption of the sorption tubes had to be conducted at 300°C. A temperature of 250°C was sufficient for desorption of arsines from the SPME fibers ( Table 9) but insufficient for arsine desorption from sorption tubes ( Appendix 5: 129).

Table 8 Optimized GC-MS conditions for analysis of volatile As on sorption tubes (Bruker mobile GC-MS EM 640)

Column DB 5 Polarity non-polar Length 12 m

Diameter 0.32 mm Film thickness 1 µm

Desorption temperature 300°C Desorption time 3 min

Injection 20s Split split valve 1 turn

Gas flow over the column ca. 1 mL/min Temperature-time-program 40/3/10, 300/10

Total time 39 min Purge 39 min

Carrier Gas N2 6.0 Carrier gas pressure 0.15 bar

MZ 70-180 Scan rate 8.2 ms/u

Scav. Time 0 min Acq. Start 0 min

As mentioned in section 2.6.1.1, HG does not always yield the volatile equivalent of the aqueous species (e.g., DMA from DMAA) only. In the experiments presented in this section and in section 3.2.3.2, five different arsine species were detected after HG from the acidified solution of dimethyl arsenic sodium salt trihydrate: MMA, DMA, TMA, methyldichloroarsine (CH3)AsCl2 and di-methylchloroarsine (CH3)2AsCl (mass spectra in Appendix 8-3, Appendix 8-5, Appendix 8-6, Appendix 8-9, Appendix 8-10). The only sorption material that was found suitable for sufficient sorption and desorption of arsines was the molecular carbon sieve Carboxen 564 ( Appendix 5: 0A-0E, 129, 131, 136). The major As species detected was TMA. Dimethylchloroarsine was found in traces only in two samples ( Appendix 5: 0A, 129). In contrast to SPME fibers (section 3.2.3.2), neither DMA nor MMA or (CH3)AsCl2 were detected on sorption tubes. Compared to Carboxen 564, the second molecular carbon sieve Carbosieve S III ( Appendix 5: 134) yielded significantly lower TMA peak areas (2-3 orders of magnitude), and the porous polymer Chromosorb 106 only half of that of Carbosieve SIII ( Appendix 5: 132). The porous polymer materials Tenax TA ( Appendix 5: 52, 53, 130), Chromosorb 105 ( Appendix 5: 133), and the graphitized carbon black Carbotrap ( Appendix 5: 135) showed no sorption / desorption at all.

Initially, sorption tubes were thought to be both cheaper and more practical in the field compared to the sensitive SPME fibers that tend to break easily. The fact that only TMA could be trapped on the Carboxen sorption tubes compared to DMA, TMA, and (CH3)2AsCl on Carboxen coated SPME


fibers (section 3.2.3.2), might indicate the higher sensitivity of SPME fibers. However, because the exact composition of the Carboxen coating on SPME fibers is patented, no information could be obtained about it. Carboxen 564 that was used for the sorption tubes in this study might not be the best choice. Carboxen 1000 with a surface area of 1200 m2/g, advertised to trap very volatile gases even without cryotrapping, Carboxen 1001 with a surface area of 500 m2/g for C1 and C2 hydro-carbons, or alternatively Carboxen 1003 with a surface area of 1000 m2/g might yield better sorp-tion and desorption characteristics. Further improvement of the method during this study was not pursued because for the field application in USA no GC-MS with a desorption port for sorption tubes was available, and the SPME technique was chosen for field applications.

3.2.3.2 Solid-phase micro-extraction (SPME)

Three different types of SPME fibers were purchased from Supelco: PDMS 100, PDMS/CAR, PDMS/CAR/DVB ( Table 5). Conditioning was done according to the manufacturer´s instructions listed in Table 5. Headspace SPME technique (section 2.7.4.2) was used and the fibers were ex-posed 15-20 minutes with the Supelco SPME holder for manual use ( Figure 25). Total volumes to headspace volumes were 300 mL to 100 mL and 26 mL to 16 mL respectively ( Appendix 7). The optimized GC-MS program is shown in Table 9.

Figure 25 Supelco SPME holder for manual use (top with fiber extended); PDMS-CAR fiber with spring for manual sampling, PDMS-CAR-DVB without spring for automatic sampling

The five arsine species MMA, DMA, TMA, (CH3)AsCl2, and (CH3)2AsCl generated from the di-methyl arsenate solution showed a very different sorption behavior on the three fiber types ( Figure 26). The pure PDMS fiber provided a very good separation of (CH3)AsCl2 and (CH3)2AsCl ( Appendix 7: 116). The arsine compounds with a lower molecular weight (MMA, DMA, TMA), however, were not sorbed at all. The separation of DMA, TMA, and (CH3)2AsCl was possible by PDMS-CAR fibers ( Appendix 7: 114, 117). The peak areas of (CH3)2AsCl were hereby about 2 orders of magnitude larger than those by sorption from PDMS. However, a plateau in the chroma-togram from 1.5 to 6.5 minutes masked the DMA and TMA peaks partly and a possible methyldi-chloroarsine (CH3)AsCl2 peak completely. Very small amounts of MMA were detected on PDMS-CAR-DVB fibers only, together with DMA and TMA ( Appendix 7: 115, 118). DMA and TMA

PDMS-CAR fiber (retracted)

PDMS-CAR-DVB fiber (retracted)

fiber extended


showed a much better separation on PDMS-CAR-DVB compared to PDMS-CAR. No chloroarsi-nes were sorbed on PDMS-CAR-DVB.

Table 9 Optimized GC-MS conditions for analysis of volatile As on SPME fibers (Bruker mobile GC-MS EM 640; in brackets conditions for GC-MS HP 5890 Series II from the Department of Analytical Chemistry Tech-nische Universität Bergakademie Freiberg; the HP 5 ms column is a low bleed column)

Column DB 5 (HP 5 ms) Polarity non-polar Length 12 m (30 m)

Diameter 0.32 mm (0.25 mm) Film thickness 1 µm (0.25 µm)

Desorption temperature 250°C Injection 20s

Split splitless 3 min, then 1:20 Gas flow over the column ca. 1 mL/min

Temperature-time-program 40/3/10, 250/10 Total time 34 min

Purge 34 min Carrier Gas N2 6.0 (He 5.0)

Carrier gas pressure 0.15 bar MZ 70-180 (70-250)

Scan rate 8.2 ms/u (2.1 / s) Scav. Time 0 min Acq. Start 0 min

The detection limits of the mobile GC-MS EM 640 proved insufficient for the detection of arsine species on PDMS 100 and PDMS-CAR-DVB fibers ( Appendix 7: 119, 128). On PDMS-CAR, the predominant arsine species (CH3)2AsCl could always be detected ( Appendix 7: 120-127). Also for TMA a slight peak was observed, however, mostly the separation from the baseline was too insig-nificant to allow for a peak area calculation. Only when a brand-new PDMS-CAR fiber was used, a clear TMA peak was detected, and the (CH3)2AsCl peak was about 2.5 times larger than for the PDMS-CAR fiber already used before ( Appendix 7: 127a; Figure 27).

This new fiber also showed much larger peaks at the end of the chromatogram than the old fiber. Those additional peaks probably are hexamethylcyclotrisiloxane [amu 96-133-(207)], octamethyl-cyclotetrasiloxane [amu 73-133-(191-207-281)], and decamethylcyclopentasiloxane [amu 73-(267-355)]. Because of the selected mass range from 70-180 some of the mass fragments of these com-pounds were not scanned (numbers in brackets), but the remaining masses and a significant correla-tion between their boiling points and retention times make the identification highly likely. Those siloxanes are a typical indication for bleeding of the sorption material, probably a slow degradation of the fiber coating. A decrease in these compounds shows an increase in degradation, and obvi-ously a decrease in sorption efficiency for arsines. This effect of larger arsine and siloxane peaks was already less pronounced for the second run with the new fiber, compared to the fiber already used about 15 times ( Appendix 7: 127b).


Figure 26 As species sorbed on different SPME fibers (experimental conditions and peak areas in Appendix 7: 116-118, analysis with GC-MS HP 5890 Series II; conditions in Table 9; mass spectra also in Appendix 8)

(CH3)AsH2

(CH3)2AsH

(CH3)3As 2

3

1

3

and 1.62 min

PDMS-CAR-DVB

PDMS-CAR

PDMS 100 (CH3)AsCl2

(CH3)2AsCl

1

3

2

5

4

4

and 2.68 min

(CH3)2AsH

(CH3)3As

(CH3)2AsCl

2

5

3

2

5

5

5

3

and 2.97 min and 3.96 min

and 1.68 min

3

4

5


Figure 27 Comparison between chromatograms from an older PDMS-CAR SPME fiber (used about 15 times; lower line in the chromatogram) and a brandnew PDMS-CAR fiber (upper line in the chromatogram), both the arsine target peaks and additional peaks probably from siloxanes are larger on the new fiber

Storing the PDMS-CAR fibers retracted in their metal shaft and wrapped in aluminum foil at 4°C in the refrigerator yielded similar peak areas as for analysis immediately after the experiment ( Appendix 7: 127b), suggesting that the arsines are stable on the fiber for at least 3 days. However, for an exact quantitative interpretation of storage losses and aging effects, the mobile Bruker GC-MS EM 640 is not suitable. For sampling in the field the quantification is further complicated by competitive sorption of other gases and unknown sorption affinities of the different arsines. Arsine can probably not be sorbed at all without using cryotrapping techniques. Thus, species-selective sampling by SPME was done qualitatively, not quantitatively.

3

1

2

4

5

PDMS-CAR (new)

PDMS-CAR (old)

(CH3)3As

(CH3)2AsCl

C6H18O3Si3 (?)

C8H24O4Si4 (?)

C10H30O5Si5 (?)

Summary of chapter 3 77


The previous sections describe the development of a robust method for in-situ sampling of volatile metallics from aqueous environments. Efficient gas-water separation is achieved with a PTFE membrane (porosity 0.1 µm) inside a specially designed PTFE collector cell that can be placed at the sediment-water interface or in the aqueous body itself. Vacuum up to -500 mbar can be applied to force the gases to a trapping system of either liquid or solid sorbents. As liquid sorbent a 1:100 diluted NaOCl solution in a PTFE bottle is used. PTFE rings in the oxidizing bottle increase the reaction surface, and thus trapping efficiency. For As, laboratory trapping efficiencies of 84% ± 6.6 were achieved. The build-up of a high pressure from the intense gas evolution during HG might be responsible for the loss of some volatile As, especially at the beginning of the reaction. In the field, continuous low-pressure arsine evolution will most likely result in even better trapping efficiencies. Quantitative evaluation of volatile metallics concentrations is achieved by referring the trapped mass in the oxidizing solution [µg] to the replaced vacuum volume [m3].

Besides this fast screening method on total volatile metallics concentrations, two species-selective sampling procedures were investigated for As. Solid phase micro extraction (SPME) fibers enabled the detection of more volatile As species in lower concentrations than solid sorbent tubes. For SPME fibers, selective sorption was observed dependent on the volatile As species. The chloro-arsines with the larger molecular weights (CH3)AsCl2 and (CH3)2AsCl were sorbed on PDMS 100 fibers whereas PDMS-CAR-DVB fibers proved best for MMA, DMA, and TMA. The PDMS-CAR fibers showed by far the largest peak areas of all fibers for DMA, TMA, and (CH3)2AsCl. The most effective sorption material in sorption tubes was also Carboxen but only TMA and traces of (CH3)2AsCl were detected. Probably the selected Carboxen type (CAR 564) does not present the optimum fit yet. Volatile As proved to be stable on the SPME fibers for at least 3 days, when wrapped in aluminum foil and stored at 4°C in the refrigerator. Quantitative evaluation could not be achieved because of insufficient detection limits of the GC-MS, unknown sorption affinities of the individual arsines on the various SPME fibers, non-availability of calibration standards, and poten-tial competitive sorption with other, unknown gases. For the field investigations in this project, presented in the following chapter 4, species-selective sampling is only used for qualitative infor-mation.

4 YELLOWSTONE NATIONAL PARK

4.1 Introduction

Yellowstone National Park is located in the northwestern corner of Wyoming, USA. It was founded in 1872 as the world´s first national park over an area of 8995 km2. It is famous for its more than 10,000 geothermal features, including 200 geysers which equals 70% of the world´s active geysers, numerous hot springs, fumaroles, and mud pots. The world´s largest geyser, Steamboat Geyser, with a maximum eruption height of more than 115 m, as well as the world´s third largest hot spring, Grand Prismatic Spring, with a diameter of approximately 110 m, are found here. The ma-jor watershed, the Continental Divide, cuts through the southwestern part of the Park. Rivers north of it discharge via the Madison River (confluence of Gibbon River and Firehole River), Gallatin River, and Yellowstone River to the Missouri and finally to the Gulf of Mexico. To the south the discharge follows mainly the Snake River and Falls River to the Pacific Ocean ( Figure 28).

Figure 28 Yellowstone National Park (dot-dashed line is the park boundary) with topographic features, riv-ers, and lakes, and an approximation of the Continental Divide (dashed line) (modified from White et al. 1988)

Continental Divide

Yellowstone Park Boundary

80 Yellowstone National Park

4.1.1 Geology of the Park

4.1.1.1 Pre-Quaternary Geology

An overview of Yellowstone´s geology is given in Smith and Siegel (2000) and a detailed technical review for the Pliocene and Quaternary can be found in Christiansen (2001). The oldest rocks found in the Park are 2.8 billion year-old granites and gneisses in the Gallatin Range at the north-central edge of the park ( Figure 29). In some locations, they are intruded by 1.5 billion year-old diabase dikes. Up through the Laramide orogeny, Yellowstone showed a similar development as the rest of the Rocky Mountain area with marine conditions through Paleozoic and early Mesozoic and the formation of broad anticlinal and overthrust structures in the early Mesozoic as part of the Basin and Range Province. Paleozoic-Mesozoic deposits outcrop only at the southern- and north-ern-most margins of the Park. The following phase of extension during early Tertiary initiated in-tensive volcanism in the Eocene. The andesitic lava flows, basalt flows, ashes and mudflows of this Absaroka Volcanic Supergroup build up the mountain ridges in the eastern part of the Park (Mount Washburn). Intrusions of igneous rocks were minor in the Tertiary and formed isolated domes east of Yellowstone Lake and in the northwestern part of the Park only (e.g., Bunsen Peak).

Figure 29 Simplified geological map of Yellowstone National Park (Kiver and Harris 1999)

Introduction 81

4.1.1.2 Quaternary Volcanism

During late Tertiary to Quaternary, Yellowstone encountered three major volcanic cycles, each of which resulted in the formation of a caldera ( Figure 30). Each explosive, pyroclastic-dominated activity was preceded and succeeded by large eruptions of rhyolitic lava. The oldest eruption oc-curred about 2.0 million years ago. It deposited 2,500 km3 of rhyolitic tuff, the so-called Huckle-berry Ridge Tuff, with maximum thicknesses of 150 to 750 m. A second, smaller cycle began about 1.3 million years ago and ended with the deposition of about 280 km3 of Mesa Falls Tuff, up to 150 m thick. The latest eruption happened 630,000 years ago. The 1,000 km3 Lava Creek Tuff covers about 60% of the Park´s surface with maximum thicknesses of 500 m. The caldera of this third volcanic cycle measures 45 km to 75 km in diameter. Compared to the 1980 eruption of Mount St. Helens, the three Yellowstone eruptions were 25000, 280, and 1000 times larger (Smith and Siegel 2000).

Figure 30 Major fracture zones and faults in and around Yellowstone National Park; I, II, and III indicate the three Yellowstone calderas formed 2, 1.3, and 0.6 Ma ago; circles picture epicenters of earthquakes with Rich-ter magnitudes between 1 and 6 (White et al. 1988)

The source of this Late Tertiary to Quaternary volcanism is the Yellowstone hotspot, the largest hot spot under a continent worldwide (Pierce and Morgan 1990, 1992; Smith and Braile 1994). The origin of this hotspot is still controversial. Murphy et al. (1998) suggested that the hotspot first appeared already more than 55 million years ago beneath the Pacific Ocean. Duncan (1982) and Pyle et al. (1997) attribute a 55 million year-old captured island chain in the coast range of Oregon and Washington to the Yellowstone hotspot. The controversy of such theories and how they relate


to the global understanding of plate movement directions, rates and major orogenic events is dis-cussed in Pierce et al. (2000). Assuming a first appearance of the hotspot 55 million years ago off the west coast, it is doubtful that there is a gap of approximately 40 million years without any sur-face manifestations of the hotspot´s track. The first clear indication of the hotspot is an old caldera (McDermitt caldera) formed 16.5 million years ago in northern Nevada near its border to Oregon and Idaho along the Nevada-Oregon rift zone ( Figure 31; Zoback and Thompson 1978). Because the North American plate drifts southwest over the Yellowstone hotspot, the eruptions formed a chain of progressively younger calderas stretching along a parabolic hotspot track from Nevada 800 km northeast up Idaho´s Snake River Plain to Yellowstone.

Figure 31 Path of the hotspot that triggered about 100 caldera eruptions from 7 major volcanic centers (McDermitt volcanic field 16.1-15.6 Ma, Owyhee volcanic center 14.7-13.8 Ma, Bruneau-Jarbidge volcanic center 12.5 Ma, Twin Falls volcanic center 11 Ma, Picabo volcanic center 10.3 Ma, Heise volcanic center 6.6- 4.3 Ma, Yellowstone volcanic center 2, 1.3, 0.6 Ma) (Kiver and Harris 1999)

The calculated average speed of the hotspot track during the last 16.5 million years was 4.5 cm annually. Until 8 million years ago, the average speed was 6.1 cm per year but then it slowed down to 3.3 cm per year. The direction of the track changed from east-northeast to northeast. Smith and Siegel (2000) explain this change in track direction as a slowdown and a change of the direction of the plate movement. In contrast, Pierce et al. (2000) state that the anomalously high average migra-tion rates during the first 8 million years are three times faster than the predicted plate movements. Pierce et al. (2000) propose that 16 million years ago the hotspot´s actual feeding tail was located 260 km east of the McDermitt caldera in a straight, continuous line with the younger calderas (no

Introduction 83

change of plate movement direction assumed). At that time, the hotspot was located underneath the northeastward-tilting subduction zone of the oceanic lithosphere of the Juan de Fuca slab under-neath the North American plate. The buoyantly rising plume head intersected with the overlying Juan de Fuca slab and was thus deflected west-southwestward, protruding to the surface at the posi-tion of today´s McDermitt caldera. Because of an increasing inclination of the Juan de Fuca slab and the continuous southwestward plate movement, the subduction zone was shifted southwest-wards away from the hotspot. Since about 10 million years ago, the plume head rises straight up to the surface from the feeding tail without any deflexions. This termination of southwestward deflex-ion can explain the apparently higher migration rates of the hotspot track to the northeast.

In the wake of the hotspot, the Snake River Plain sank as much as 600 m and was flooded by a series of basalt lava flows that today cover the rhyolites from the explosive eruptions and smoothed the topography of the old calderas. In Yellowstone, at least 30 subsequent smaller lava flows filled the youngest caldera from the third major volcanic cycle. The most recent lava flow, the Pitchstone Plateau flow in the southwestern part of the Park, dates back about 70,000 years (Smith and Siegel 2000).

Several recent reports argue against the theory of a Yellowstone deep mantle plume. In contrast to hotspots underneath oceanic plates, the Yellowstone hotspot probably does not root at the mantle-core boundary, but begins as a melting anomaly in the upper mantle at about 200 km beneath the surface (Christiansen et al. 2002; Humphreys et al. 2000). Seismic data show a low-wave-speed layer in this depth, extending no deeper than 200-250 km (Christiansen et al. 2002; Iyer et al. 1981). An extending region over a melting anomaly would not only explain the propagation of the Yellowstone hotspot to the northeast, but also another melting anomaly to the northwest, the New-berry anomaly.

4.1.1.3 Quaternary Glaciation

In addition to the filling by younger lava flows, the topography of the Yellowstone calderas was smoothed by glaciation. The Yellowstone plateau with average altitudes of about 2000 to 2500 m was covered by ice up to 1200 m thick during at least two major ice stages: Bull Lake stage from 160,000 to 130,000 years ago, and Pinedale stage from 50,000 to 14,000 years ago (Pierce 1979). These glacial events overlap with the volcanic events during Quaternary.

During glaciation the potentiometric surfaces for groundwater in the underlying rocks were raised hundreds of meters above the ground surface, raising the boiling point curves (Bargar and Fournier 1988). Excess thermal energy became stored in rocks at shallow depths. When the glaciers receded, this excess energy was released enhancing the geothermal circulation. Rapid draining of local lakes lowered the hydrostatic head to such an extent that pore water flashed to steam and produced sev-eral large hydrothermal explosion features, like e.g., Pocket Basin in Lower Geyser Basin or Roar-ing Mountain (Muffler et al. 1971).


4.1.1.4 Hotspot development after the last volcanic cycle

The hotspot under Yellowstone´s youngest caldera consists of a column of hot rock with about 10-30% molten rocks, and extends upward to about 80 km beneath the surface. Basalt blobs are sheared off the top of the hotspot by the movement of the lithospheric plates. As these blobs rise in the crust, they melt silica-rich rock, to create a large sponge-like chamber of partly molten rhyolite. This magma chamber extends from about 8 to 13 km beneath the surface. In the northeastern part of the Park it might be as shallow as 5 km beneath the surface (Smith and Siegel 2000).

After the last volcanic cycle, two main conduits remained connected to the magma chamber above the Yellowstone hotspot. At the surface, they are expressed as two resurgent domes with a south-east-northwest striking ( Figure 30). The 630,000 year-old Sour Creek dome has an extension of 10 x 16 km and rises 400 m above the surrounding countryside north of Yellowstone Lake. About 32 km southwest of Sour Creek, north of the Upper Geyser Basin, the 150,000 year-old Mallard Lake dome rises 300 m above its surroundings with a size of 11 x 8 km. Both domes are interconnected by the 153,000 year-old Elephant Back fault zone (Obradovich 1992) that outcrops for about 6 miles in between the two domes reaching altitudes of more than 2600 m (Smith and Siegel 2000).

Ever since the last volcanic cycle, Yellowstone caldera has been subject to uplift and subsidence. For the last 12,000 years this movement is expressed in deformations on postglacial terraces around Yellowstone Lake (Locke and Meyer 1994; Pierce et al. 2002) and changes in the gradient of the Yellowstone river (Pierce et al. 1997). Pelton and Smith (1982) documented an uplift of 14 ± 1 mm/year of the caldera floor in the 20th century by comparing leveling surveys from 1923 and 1975-1977. Further investigations by leveling and differential GPS measurements along a traverse from Lake Butte to Canyon Junction (and from 1989 further on to Mount Washburn) documented an uplift of the northeastern caldera floor (center at Le Hardy Rapids) from 1976 to 1984 with an average rate of 22 ± 1 mm/year. From 1984 to 1985, a phase without measurable deformation fol-lowed. Subsidence started in 1985 and proceeded through 1995 with an overall rate of -19 ± 1 mm/year. The rate was higher (-39 ± 5 mm/year) in the years 1993 to 1995 only (Dzurisin and Yamashita 1987; Dzurisin et al. 1990, 1994, 1999). From 1995 to 1998, a leveling survey by Dzurisin et al. (1999) showed a renewed uplift of 8 ± 1 mm/year.

Since 1992, high resolution synthetic aperture radar (SAR) images obtained from European Space Agency satellites ERS-1 and ERS-2 were used to measure year-to-year deformation over the whole surface of the caldera instead of movement on isolated points only (Dzurisin et al. 1999; Wicks et al. 1998). The resultant interferograms enable a much more detailed location of deformation sources. Dzurisin et al. (1999) and Wicks et al. (1998) observed that deformation centers changed on a scale of 1 to 2 years between the two resurgent domes. The center of subsidence shifted from Sour Creek dome during August 1992 to June 1993 to Mallard Lake dome during June 1993 to August 1995. Between August 1995 and September 1996, while the Mallard Lake dome continued to subside slightly, the region near the Sour Creek dome began to inflate. From September 1996 to June 1997, uplift proceeded at Mallard Lake dome. By June 1997, uplift extended across most of the caldera. Its center had risen by approximately 40 mm since August 1995 (Dzurisin et al. 1999).

Introduction 85

The initiation of the renewed uplift from Sour Creek dome to Mallard Lake dome may indicate that the driving pressure source lies under Sour Creek dome or that Sour Creek dome has a preferential connection to the magma chamber below. A second conduit probably interconnects the two domes. Wicks et al. (1998) suggest a third conduit where fluids flow out of the caldera, initiating subsi-dence of the whole area. This third conduit is believed to extend to the northwestern corner of the park, feeding the geothermal features of Norris Geyser basin, and maybe connected to the Norris-Mammoth corridor ( Figure 30). The Norris-Mammoth corridor is a complex subsidence structure of faults, volcanic vents, and thermal activity that extends north from Norris at least up to Mam-moth, probably even further up to Corvin Springs. This structure was inferred from electromagnetic (Stanley et al. 1991), seismic, and gravity data (Lehmann et al. 1982). The hypothesis of this struc-tural conduit is supported by the location of the two largest earthquake swarms ever recorded in the park since 1972 west of Madison junction, just outside of the northwest caldera rim. They initiated or accompanied the changes from uplift to subsidence in 1985 and to renewed uplift in 1995. The 1985 earthquake swarm started in October and persisted through 1986 with 28 earthquakes of a magnitude > 3.5, the strongest being 4.9. The 1995 earthquake swarm began in late June and con-tinued through mid-July with more than 560 earthquakes. The largest earthquake had a magnitude of 3.1 (Dzurisin et al. 1999). Most recent research reveals an inflation area of 30 x 40 km at the center of the proposed outlet (Wicks et al. 2002, 2003). From 1996 to 2002, a maximum uplift of 125 mm was detected over this Norris Uplift Anomaly.

The primary cause of uplift and subsidence in the two main resurgent domes may be explained by two end-member models (Dzurisin et al. 1990; Fournier 1989). Initiation of an uplift may be caused by the injection of basalt near the base of the cooling rhyolitic magma chamber. In the up-per part of the magma chamber, rhyolite crystallizes and releases magmatic fluids in the shallow hydrothermal system. Subsidence starts, when the basalt supply rate is less than the subsidence caused by crystallization and loss of fluids. The thermal energy discharged by the advecting ther-mal waters could be supplied by the latent heat of crystallization of about 0.2 km3 rhyolitic magma per year that would lead to a volumetric contraction of about 0.014 km3. That contraction is close to the observed rate of subsidence from 1985 to 1995. The second model assumes that uplift is pri-marily caused by pressurization through magmatic fluids released during crystallization and trapped at lithostatic pressure beneath an impermeable self-sealed zone of mineral deposition. Sub-sidence occurs during episodic hydrofracturing. The short intervals of change in subsidence and uplift between the two resurgent domes (1-2 years) favor the second hypothesis with a low-viscosity fluid (gas plus brine) over a more viscous one (magma). Assuming that the magmatic fluids released from the crystallization of 0.2 km3 rhyolitic magma per year contain about 2 wt% water, and are trapped at lithostatic pressure at a depth of 4-5 km that results in a net volume in-crease of about 0.026 km3/year that can easily account for the observed inflation from 1923 to 1984 (Fournier 1989).


4.1.1.5 Hydrogeochemistry of geothermal features

Surface geothermal activity expressed in hot springs, geysers, fumaroles and mud pots indicates the final stage of the third caldera cycle. It started about 15,800 years ago after the Pinedale glaciation (Fournier et al. 1976). Most surficial geothermal features occur along the caldera´s ring fracture zone and along the Norris-Mammoth corridor ( Figure 32) in flat-bottomed valleys between lava flows with permeable stream and glacier sediments that allow water to percolate into the ground.

Figure 32 Yellowstone NP with the youngest caldera rim and ring-fracture zone, and the resurgent domes Mallard Lake (southwest), and Sour Creek (northeast); black areas indicate active and fossil geothermal systems (White et al. 1988)

Mainly over the last 40 years, the hydrogeochemistry of the geothermal features has been investi-gated intensively (Allen and Day 1935; Ball et al. 1998a, 1998b, 2001, 2002; Fournier 1989; Fournier et al. 1992, 1994a, 2002; Rowe et al. 1973; Thompson et al. 1975; Thompson and Yadav

Introduction 87

1979; Thompson and DeMonge 1996; Truesdell and Fournier 1976; Truesdell et al. 1977; Stauffer et al. 1980). Several geochemical models were derived.

In the north and northwest where interlayered permeable rhyolite flows from the third and second caldera cycle form possible fluid reservoirs and water recharges from the Gallatin Range, there is a concentration of larger geothermal areas with geyser and hot springs discharging mostly high Cl and SiO2 waters at a near-neutral pH (Truesdell et al. 1977). The eastern part of the Park, where impermeable ash flow tuffs predominate, is dominated by geothermal features with little or no dis-charge, such as mud pots and fumaroles (White et al. 1971). Stable isotope studies revealed that more than 95%, maybe even as much as 99.6-99.9% of the geothermal water circulating in the Yellowstone caldera is meteoric water, mixed with only little magmatic fluids (Craig et al. 1956; Fournier 1989). Infiltrating water is heated to about 350°C (inferred from enthalpy-chloride rela-tions and Si-geothermometer) over the magmatic heat source in 4-5 km depth. This “parent water” contains about 400 mg/L Cl, high Na concentrations, high gas concentrations (mainly CO2, and H2S), and low concentrations of SO4

2- and HCO3- (Fournier et al. 1986; Fournier 1989).

As geothermal fluids ascend adiabatically from the magma chamber, water and steam separate as pressure decreases at shallower depths, and at fractures or fissures. Two major water types form:

• Steam, rich in CO2, and H2S, rises and condenses or mixes with shallow groundwater. The H2S is oxidized abiotically to elemental sulfur and sulfur is oxidized biotically to sulfuric acid. This water type (type I) is characterized by low pH, and high concentrations of S(VI) and high PCO2. Thiosulfate concentrations are low due to the instability of S2O3

2- in acid wa-ters (Xu and Schoonen 1995; Xu et al. 1998). Increased concentrations of Al (on average several mg/L up to a maximum of 100 mg/L; Ball et al. 2002) and rare earth elements (20-1130 nmol/kg Σ REE; Lewis et al. 1997) are found from reactions with surrounding rhyolite.

• The residual liquid phase, depleted of the gaseous compounds separated with the steam, forms water type II. This water type has a near-neutral pH, and lower concentrations of S(VI) and lower PCO2 compared to type I waters. The elements Si and Cl are enriched, often showing higher concentrations than the 400 mg Cl/L in the parent water. Rare earth element concen-trations in type II waters with a near-neutral pH are mostly at or below detection limits (Lewis et al. 1997).

Both type I and type II waters can be diluted near or at the surface by meteoric waters. Mixing ra-tios can be determined using 2H and 18O isotopes (Ball and Nordstrom 2001; Kharaka et al. 1991, 2002; Truesdell et al. 1977). From mixing with meteoric waters, the Cl/S(VI) ratio of type II waters decreases. In these waters of slightly acidic to near-neutral pH, and enriched in S(VI) (Cl/S(VI) ratio 10-25), S2O3

2- may exceed dissolved H2S concentrations. As a result of subsurface H2S oxida-tion, elemental sulfur hydrolysis, or interaction with sulfate-bearing minerals, S2O3

2- concentrations of 9 to 95 µmol/L are found (Xu et al. 1998). The 3He/4He ratios of the geothermal waters are typi-cally increased compared to atmospheric values, indicating a mantle source. Especially high values occur in water that discharges directly from deep reservoirs, lower values in waters depleted in mantle gas (Kennedy et al. 1985; Kharaka et al. 1991, 2000, 2002).


Arsenic is a typical element that occurs at high concentrations in geothermal fluids (Webster and Nordstrom 2003). At depth, most reservoir fluids are undersaturated with regard to As minerals. High As concentrations in geothermal features most likely result from leaching of the surrounding rocks during the ascent of the fluids as shown by Ellis and Mahon (1964) and Ewers and Keays (1977) using rock-leaching experiments. Vapor-liquid distribution factors for As (Km = mAs(vapor)

[mol/kg] / mAs(liquid) [mol/kg]) measured in different hot springs of modern geothermal fields typi-cally are in the range of 0.001 to 0.01 at temperatures of 150 to 300°C (Ballantyne and Moore 1988). Experimental studies by Pokrovski et al. (2002) revealed that the small fraction of As that is separated into the steam (= type I waters at Yellowstone) is mainly As(OH)3(g). In the presence of HCl(g) and at temperatures <200°C, AsCl3(g) might predominate. Gaseous As sulfides were found to be negligible in the presence of H2S. Type II waters contain higher As concentrations compared to type I, since the majority of As remains in the liquid phase during phase separation (Stauffer and Thompson 1984). This behavior is similar to Cl and the reason for the positive correlation of As and Cl in geothermal fluids, first observed by Ritchie (1961). In Yellowstone, many hot springs are enriched in As relative to Cl. This enrichment is either caused by high CO2 concentrations increas-ing arsenite hydrolysis and solubility from host rocks (Stauffer and Thompson 1984) or by accumu-lation of As in precipitates near the surface (Nordstrom et al. 2001). A review on the source, trans-port, and fate of geothermal As is provided in Webster and Nordstrom (2003).

In Yellowstone, As was first detected by Gooch and Whitfield (1888). Concentrations range from 0.1 to 9 mg/L (Ball et al. 1998a, 1998b, 2001, 2002; Nordstrom et al. 2001; Stauffer and Thompson 1984). The As(III)/As(V) ratios are extremely variable with a distinct bimodal distribution of As being either fully reduced in hot springs or fully oxidized in drainages (Nordstrom et al. 2001). Upon discharge from the hot spring, As(III) oxidizes rapidly to As(V) (Stauffer et al. 1980; Stauffer and Thompson 1984). Langner et al. (2001) determined a half-life of 0.58 minutes for the microbially mediated oxidation at a hot spring in One Hundred Springs Plain, Norris Geyser Basin (“Dragon Spring”, pH 3.1, temperature 58-62°C). Assignment of specific microbial populations to the As(III) oxidation and detection of their oxidation mechanisms - chemolithoautotrophic metabo-lism or extracellular oxidative detoxification - is under research. Often, a distinct zoning of yellow, brown and green microbial mats with increasing distance from a hot spring discharge is observed. Jackson et al. (2001) determined for “Dragon Spring” that 90% of the bacteria in yellow and brown microbial mats are Hydrogenobacter acidophilus (H2 or S0 oxidizing bacteria) and Desulfurella sp. (S0 reducer). A strain of the arsenite-oxidizing Hydrogenobaculum was later isolated from the same spring by Donahoe-Christiansen et al. (2004). Archaeobacter were found in the brown mats only, Cyanidium caldarium was identified in the green mats (Langner et al. 2001). Investigating Twin Butte Vista hot spring in the Lower Geyser Basin (pH 8.8, temperature 65-82°C), Gihring et al. (2001) found that Thermus aquaticus oxidizes As(III) rapidly. Laboratory experiments confirmed that Thermus species are not able to grow with As as the only energy source, and no catabolic en-ergy is derived from As(III) oxidation. Thus, As(III) oxidation is likely a detoxification mechanism for this species.

Introduction 89

No investigations have ever been conducted in Yellowstone on methylated or volatile As com-pounds before the current project. In general, only few studies exist about organic and volatile As speciation in geothermal gases or waters. Crecelius et al. (1976) detected volatile As compounds in non-condensable gases in several hot-water fields in California. Total concentrations reached up to 79 ng/kg. No As speciation was performed. Hirner et al. (1998) detected AsH3, MMA, DMA, and TMA in geothermal gases over hot springs in British Columbia. However, no quantification was achieved. In geothermal waters in New Zealand, methylated As made up 0.02-8.9 µg/L and an unidentified, probably organic As compound up to 30 µg/L (Hirner et al. 1998; see also Table 1).

4.1.2 Study areas

Five study sites were selected for hydrogeochemical sampling with focus on dissolved and volatile As speciation. From the north to the south, those areas are:

• Nymph Lake area with Frying Pan Spring • Hazle Lake area • Ragged Hills area in the Norris Geyser Basin • Geyser Springs group in the Gibbon Geyser Basin • Pocket Basin in the Lower Geyser Basin

Only Pocket Basin lies within the Yellowstone caldera. The areas 1-4 are located to the north of the caldera, along the Norris-Mammoth corridor (section 4.1.1.4; Figure 33).

White et al. (1988) present a model for the hydrogeochemistry of hot springs along the Norris-Mammoth corridor ( Figure 34). A magma chamber is hypothesized to be centered near the vent of the 180,000 years old Obsidian Cliff flow. From there, geothermal fluids are flowing to the north and to the south.

To the north, extensive cooling and dilution by meteoric water occurs. The waters dissolve carbon-ate from pre-Tertiary sediments that is finally precipitated upon discharge at Mammoth hot springs. Kharaka et al. (1991) propose an alternative model with a separate magmatic body for the Mam-moth hot springs area. This model is supported by high 3He/4He ratios in the geothermal waters at Norris and Mammoth and low 3He/4He ratios in all the major geothermal areas in between. The high 3He/4He ratios indicate a direct release from a deep reservoir. Additional evidence for a mag-matic body in the Mammoth area comes from magnetotelluric and other geophysical surveys, indi-cating a partially molten area south of Bunsen Peak at a depth of 6 km (Stanley et al. 1991).


Figure 33 Geological map for the study sites; outside the caldera, along the Norris-Mammoth corridor: 1 = Nymph Lake (incl. Frying Pan Springs about 300 m south; Figure 35), 2 = Hazle Lake ( Figure 36), 3 = Ragged Hills (Norris Geyser Basin; Figure 39), 4 = Geyser Springs group (Gibbon Gey-ser Basin; Figure 40); inside the caldera: 5 = Pocket Basin (Lower Geyser Basin; Figure 41); shaded areas show named Quaternary rhyolitic lava flows and domes, shading in-dicates age according to legend (modified after Fournier et al. 1994b)

1 2

3

5

45km 1 2 3 4 0

Introduction 91

In the model of White et al. (1988), the hot Cl- and SiO2-rich waters of neutral pH flowing to the south, undergo subsurface boiling with a steam-water separation. At temperatures about 370°C, the steam rises and forms acid-altered areas by near-surface oxidation of steam-enriched H2S. Accord-ing to high resolution aeromagnetic mapping by Finn and Morgan (2002), there are many more yet unrecognized areas of hydrothermal alteration in the shallow subsurface along the Norris-Mammoth corridor. The residual water phase, enriched in chloride, discharges further to the south in the Norris Geyser Basin at temperatures around 270°C (section 4.1.2.3). The occurrence of acid alteration areas north of Norris is hard to combine with the detected low 3He/4He ratios in all the major geothermal areas between Mammoth and Norris (Kharaka et al. 2000). An early steam sepa-ration from a magmatic source in the Obsidian Cliff area explains the low pH, but it should yield higher 3He/4He ratios. More recent data summarized by Fournier (2004) indicate that either an in-jection of magma or an injection of hydrothermal fluid (or both) from the caldera magma has moved along faults to the Norris Basin and to Mammoth.

Figure 34 Model to explain hydrogeochemistry along the Norris-Mammoth corridor (White et al. 1988); from a central magma chamber underneath the area of Obsidian cliff geothermal fluids flow north to Mammoth (exten-sive cooling and dilution by meteoric water, dissolution of carbonate pre-Tertiary sediments) and south to Norris (steam-water separation; formation of acid-altered areas by near-surface oxidation of steam-enriched H2S closer to the magma chamber, discharge of residual chloride enriched water in the Norris Geyser Basin); numbers indi-cate study areas (1 = Nymph Lake, 2 = Hazle Lake, 3 = Norris Geyser Basin, 4 = Gibbon Geyser Basin, south of Norris dome); Norris dome is the third resurgent dome (Norris Uplift Anomaly) mentioned in section 4.1.1.4

4.1.2.1 Nymph Lake area

Nymph Lake is located 3.5 km northwest of Norris. The lake stretches NE-SW with a maximum length of 500 m, and a maximum width of 150 m ( Figure 35). Two geothermal areas with numer-ous fumaroles and hot springs ranging from centimeters to a few meters in size border the lake to the northeast and west. During the winter of 2002/2003, a new geothermal feature broke out, mark-ing a line of fumarole vents in the lodgepole pine forest above the western geothermal area.

Nymph Lake has one major and two minor tributaries apart from the discharge of the geothermal areas. The northern tributary rises about 2 km northwest in the Twin Lakes. The northern lake of the Twin Lakes receives some drainage from Roaring Mountain. The second tributary on the north-eastern side of Nymph Lake is the 170 m long, 1-2 m wide, and 1-10 cm deep acid Nymph Creek

insert see Figure 37

1 2

3 4

S N


originating from multiple springs (pH approximately 2.8, temperature 45°C to 62°C; Ball et al. 2002). Its distinct green color stems from the growth of 2-3 mm thick mats of the acidophilic alga Cyanidium caldarium. Although the alga´s color is exactly the same as that of many prokaryotic blue-green algae, careful scrutiny of the cells showed that it belongs to the group of eukaryotic red algae (Ferris et al. 2003; Henson and Ferris 2003). About 50 m south of Nymph Creek, “Frying Pan Spring Creek” flows into Nymph Lake, draining the Frying Pan Spring area about 250 m southeast. Note that names in quotation marks used here and in the following are unofficial names, most of them adapted from the research group of Dr. Kirk Nordstrom, United States Geological Survey in Boulder, Colorado.

Figure 35 Nymph Lake area in-cluding Frying Pan Spring in the southeast; numbers 1-6 indicate sampling sites YNPNL01-YNPNL06

Because of the preponderance of gas discharge related to the model presented in Figure 34, the hot springs at Frying Pan Spring area show high concentrations of S(VI) (440 mg/L), S2O3

2- (0.6 mg/L), and dissolved H2S (1 mg/L; Ball et al. 1998b). Ferris et al. (2003) and Ball et al. (1998b) determined 0.07 to 0.2 mg/L and <0.001 to 0.12 mg/L H2S respectively for the source water of Nymph Creek and attribute the higher concentrations downstream mainly to microbial reduction within the microbial mat. Goff and Janik (2002) determined 1.074 mol% dry gas H2S at Frying Pan Spring, compared to 0.677 mol% at Beryl Spring, 0.387 mol% at Fountain Paint Pots, and 0.001-

Nymph Lake

Frying Pan Spring

Nymph Creek

Mammoth, Twin Lakes

Norris

4 3

2

1

6

5

line of 2003 fumaroles

N0 40 80 120 160 200 m

hydrothermal alteration area

shallow lake area

wetland II II II II II II II

Introduction 93

0.298 mol% in the Mud Volcano area. Only fumaroles in the Lamar region showed higher values with 1.2-1.7 mol%. The concentration of H2 (0.3305 mol%) is also 1-3 orders of magnitude higher than in other areas in the Park. Total dissolved As concentrations of 0.1-0.2 mg/L are significantly lower than in Norris Geyser Basin, Gibbon Geyser Basin or Lower Geyser Basin (Ball et al. 1998b).

4.1.2.2 Hazle Lake area

Hazle Lake area lies about 1 km southeast of Nymph Lake area, 250 m east of the road Norris-Mammoth ( Figure 36). It consists of the lake itself with a length of approximately 600 m NNE-SSW and a maximum width of 200 m, and an adjacent wetland area with a length of 1 km, and a maximum width in the northeast of 700 m. In the northeast there is a 150 m by 400 m large zone of shallow water with significant thermal activity and precipitated alteration products. Gas discharges in shallow water can also be observed all along the lake´s western side. Several geothermal areas, mainly on the northern and western side, discharge hot water into the lake.

Figure 36 Hazle Lake area with periodic discharge to the south into the Gib-bon river; numbers 1-3 indicate sampling sites YNPHL01-YNPHL03

Apart from the geothermal discharge, the lake is fed by one tributary rising about 1 km northeast of the Hazle Lake wetland area. The lake discharge flows at first to the west, then to the south and

0 100 200 300 400 500 m N

Mammoth, Nymph Lake

Norris

Hazle Lake


shallow lake area

wetland II II II II II II II

periodical discharge to Gibbon river

3

1

2


joins Gibbon river. Inflow, outflow and lake level vary seasonally. Little research has been con-ducted at Hazle Lake so far. In 2000, U.S. Geological Survey researchers found significantly in-creased concentrations of inorganic and methyl-mercury ( Table 10). A joint project between the U.S. Department of Energy's Idaho National Engineering and Environmental Laboratory, the U.S. Geological Survey and the University of Nevada-Reno in autumn 2003, revealed significant mer-cury emissions of 200-700 ng/m2/h all along the Norris-Mammoth corridor, with a maximum con-centration of 2,400 ng/m2/h at Roaring Mountain (Department of Energy 2003).

Table 10 Unpublished data for Hg and Methyl-Hg from USGS researchers

Site Date Total Hg [ng/L] Methyl-Hg [ng/L] Hazle Lake Shore #1 9/24/2000 300 0.01 Hazle Lake Shore #2 9/24/2000 345 0.25 Hazle Lake Shore #3 9/24/2000 310 0.02

Hazle Lake inlet 9/24/2000 100 1.70 Hazle Lake outlet #1 9/24/2000 320 0.55 Hazle Lake outlet #2 9/24/2000 270 0.95

Hazle Lake outlet of wetland 9/24/2000 115 2.10

4.1.2.3 Norris Geyser Basin

Norris Geyser Basin, 4 km north of the northwest rim of the Yellowstone caldera, shows the great-est variety of geothermal features within the Park. It is located over the intersection of the Norris-Mammoth corridor and the eastward extension of the Hebgen Lake fault system. The Hebgen Lake earthquake on August, 17th 1959, with a magnitude of 7.5 was the strongest earthquake in recent times in the immediate vicinity of the Park (Marler 1964).

The hydrogeochemistry of the Norris Geyser Basin has been investigated and described among others by Ball et al. (2001, 2002) and Fournier et al. (1986, 1992, 1994a, 2002). As mentioned in section 4.1.2, Norris Geyser Basin receives the residual water phase from subsurface boiling of geothermal waters rising from the hypothesized Obsidian Cliff magma chamber ( Figure 34). In an enlarged section of Figure 34, Figure 37 shows the hydrogeochemical model derived for Norris Geyser Basin from White et al. (1988). Near-neutral pH waters high in Cl and SiO2, depleted in CO2, H2S, and other gases discharge along the central belt of springs and geysers without substan-tial dilution by cold water before boiling. Where near-surface mixing with meteoric waters in-creases S(VI) compared to Cl concentrations, 5-8 µmol/L S2O3

2- may be found (Xu et al. 1998). Acid waters occur in two belts along the northwestern and southeastern margin of the Norris Gey-ser Basin, and partly also in the central belt. White et al. (1988) propose that the acid waters result from surface acidification caused by oxidation of H2S and relative enrichment due to evaporation. Oxidation of elemental sulfur is assumed at least for the eastern belt with a local circulation of wa-ter from the Norris dome. Reservoir temperatures underneath Norris Geyser Basin of 270°C to more than 300°C are higher than in other areas in the Park.

Norris Geyser Basin undergoes an annual basin-wide hydrothermal disturbance that occurs com-monly in late summer (August / September). It is expressed in small hydrothermal explosions, a sudden increased discharge of thermal water and steam, an increased turbidity, and extreme fluc-

Introduction 95

tuations in temperature of many hot springs. Increased boiling also leads to enrichment in dissolved constituents in many hot springs. Observations by Fournier et al. (2002) during the seasonal distur-bance in 1995 revealed that the disturbances only affected the hot springs from relatively deep res-ervoirs in the central belt. The chloride hot springs with near-neutral pH picked up an acid-sulfate component and became isotopically heavier for several days after the onset of the disturbance, in-dicating subsurface mixing. Initially acid waters and near-neutral pH waters from shallower reser-voirs showed little response to the disturbance.

Figure 37 Schematic model of water types at Norris Geyser Basin (White et al. 1988)

According to models from different research groups, several factors trigger the seasonal distur-bance. A seasonal scarcity in meteoric water results in minor dilution and consequently supersatu-ration of geothermal fluids especially with respect to silica minerals (White et al. 1988). High deposition rates lead to an increased rate of self-sealing. This self-sealed silica layer that can be identified in most of the research drill holes in the Park significantly decreases permeability (Dob-son et al. 2003), increases the build-up of overpressure and finally leads to hydrofracturing fol-lowed by small hydrothermal explosions. Flow channels clogged by clay mineral deposition might produce similar effects. Deformation and fracturing induced by tectonic forces are additional, less frequent processes responsible for pressure changes within the hydrothermal system. Yet, the most important seasonal trigger is the fluctuation of the local potentiometric surface. A lower water table in summer leads to a decreasing pressure from the overlying cold water column. The boiling point within the hydrothermal system drops. In deeper, high-temperature reservoirs, excess heat released from previously heated rock causes evaporative boiling of ascending fluids. Eventually, the pres-sure at greater depths decreases, since the rate of water expulsion exceeds its supply from the deeper reservoirs. With the decline in pressure, water from shallow reservoirs, where temperatures


are less than the prevailing boiling temperature, can enter the discharge vents without hydrothermal explosions. There they mix with the waters from the deeper reservoirs, producing the change in chemistry of some hot springs described in the previous paragraph. The higher temperature gradi-ents at Norris Geyser Basin with reservoir temperatures of 270°C to more than 300°C compared to lower-temperature gradients at shallower reservoirs as e.g., Lower Geyser Basin (200-125°C) might be the reason that the annual disturbance phenomena was only observed at Norris. It might also be a lack of the typical indicator, the distinct turbidity of hot springs at the beginning of the disturbance. This turbidity is owed to clay from acid altered rocks that are widely distributed at Norris but scarce at other Geyser Basins (Fournier et al. 2002).

Ragged Hills, the area investigated within the Norris Geyser Basin, is located about 1 km west of the Norris parking lot ( Figure 38). It is one of the most recent and most dynamic geothermal areas in the Park with an anomalously high total heat flow.

Figure 38 Overview Norris Geyser Basin with Ragged Hills area; numbers 3, 4, 9, 26, and 27 indicate sam-pling sites YNPRH03, 04 (“Verde Pool” located in a hydrothermal explosion crater), YNPRH09, YNPRH26 (“Dragon Spring”), YNPRH27 (“Steam Engine”); dotted lines = 7500 ft altitude; dashed lines = boardwalk through Back Basin and Porcelain Basin; inset The Gap ( Figure 39)

0 110 220 330 440 550 m N

Ragged Hills

periodical discharge from Hazle Lake area

26

Gibbon River

The Reservoir

3

4

Norris museum

parking lot

Back Basin

Porcelain Basin

Sieve Lake

Nuphar Lake

Tantalus Creek

9

Madison Junction

Mammoth

Canyon

Gibbon River

discharge from Nymph Lake area

Gibbon River

27 Figure 39

One Hundred Spring Plain

Introduction 97

The Ragged Hills are low hills of hydrothermally cemented older Pinedale kame deposits. Their crude stratification, chaotic bedding and cross stratification, lateral discontinuity of bedding and slump structures are typical for their formation beneath a glacier by melting of the ice due to hot spring activity. A detailed geological overview including a map of a scale of approximately 1:2,300 is provided by White et al. (1988).

In the northern part of the Ragged Hills area, about 1.2 km west of the Norris parking lot, and 400 m east of the Gibbon river, a crater-like structure with a rim of Lava Creek Tuff is interpreted as a hydrothermal explosion crater thought to have formed during the waning stage of the last glaciation (Muffler et al. 1971; White et al. 1988; Figure 38). Rapid draining of a local lake lowered the hy-drostatic head to such an extent that pore water flashed to steam, producing an explosion. Today, the drainageless structure contains hot water of greenish-turquoise color, “Verde Pool”.

Figure 39 Ragged Hills, The Gap (insert of Figure 38; modified from Becker 2004); numbers 1, 2, 5-8, 10-25, and 28 indicate sampling sites YNPRH 01 (“Milky Way”), YNPRH02, YNPRH05-07, YNPRH08 (Crystal Springs), YNPRH10 (“Rainbow Growler”), YNPRH11-12, YNPRH13 (“Elk Geyser”), YNPRH14-16, YNPRH17 (“Dynamo”), YNPRH18 (“Kaolin Spring”), YNPRH19 (“Persnickety Geyser”), YNPRH20 (“Titanic Spring”), YNPRH21 (“Lifeboat Spring”), YNPRH22-25, YNPRH28 (“Milky Way II”); dashed lines = 7500 ft altitude

1

2

6

7 8

10

5

12

13

14

11 15

20

17

18

19

21

16

22

23

24

25

28

Ragged Hills

Verde Pool

N0 18 36 54 72 90 m

wet zones

discharges

hot springs

sites of intensive degassing activities


In a small area south of “Verde Pool” called The Gap an increase in temperatures was recognized from 1992 to 1995. Small seeps and springs appeared before intense hydrothermal activity began in 1995 with several eruptions ( Figure 38, Figure 39). The western part of The Gap is drained directly towards Gibbon river. The drainage of the eastern part joins Tantalus Creek in the One Hundred Spring Plain before finally discharging to Gibbon river, approximately 1 km further north than the western drainage. In 1995, three springs erupted in an area of 20 by 50 m (Ball et al. 2002). Several later eruptions especially in 1999 led to a rapid growth of the thermal area that measured 270 m by 2150 m in the year 2000. The increased activity in 1999 correlates with a peak in the annual num-ber of earthquakes (3,172 compared to less than 1,500 in the preceding four years). Ball et al. (2002) photo-documented significant changes and an increase in geothermal activity in The Gap from 1999 to 2001. “Persnickety Geyser” was a separate geyser in 1999 but now overlaps with “Titanic Spring”. The small land bridge between “Titanic Spring” and “Lifeboat Spring” subsided. During the winter 2002/2003 a discharge breakthrough formed from “Kaolin Spring” to “Milky Way”. The hydrogeochemistry of the waters in the whole Ragged Hills area covers a wide range from acid to near-neutral pH, from a predominance of S(VI) to Cl, and As concentrations of 0.2 to 9 mg/L (Ball et al. 2001, 2002).

4.1.2.4 Gibbon Geyser Basin

Geothermal activity in the Gibbon Geyser Basin is attributed to the Norris dome near the vent dome of the 90,000 year-old Gibbon river flow (White et al. 1988; Figure 34). Gibbon Geyser Ba-sin comprises the geothermal areas of Gibbon Hill Geyser in the north, Sylvan Springs west of the road Norris-Madison Junction, and Artist Paint Pots and Geyser Springs group east of the road. Geyser Springs group that was investigated in this study lies 1.3 km east of the road Norris-Madison Junction, 0.6 km southeast of Artist Paint Pots, at the eastern side of the 2455 m (8054 ft) high Paintpot Hill ( Figure 40). The whole geothermal area is approximately 600 m long, and at the maximum 150 m wide. The hot springs are distributed over four levels at altitudes of about 2286 m (7500 ft), 2298 m (7540 ft), 2310 m (7580 ft) and 2316 - 2347 m (7600 - 7700 ft). The most promi-nent feature is Avalanche Geyser at the southeastern most tip of the first level with an eruption interval of about 9-10 minutes and eruption durations of 2.5-3 minutes. The Geyser Springs group area is drained by Geyser creek that rises 1.8 km SSE from Avalanche Geyser, and discharges to-gether with the discharge from Gibbon Hill Geyser group into the Gibbon river. Both acid, sulfate-rich and near-neutral pH, chloride-rich hot springs are present at the Geyser Springs group. Arsenic concentrations range from about 0.1-0.3 mg/L in the acid springs to 1.9-2.4 mg/L in the near-neutral pH springs (Ball et al. 1998a, 2001).

Introduction 99

Figure 40 Geyser Springs group in the Gibbon Geyser Basin; numbers 1-8 indicate sampling sites YNPGG01 - YNPGG08; 7600, 7800 and 8000 are altitudes in feet; A = Avalanche Geyser; O = Oblique Geyser; 6 = “Bull´s Eye”

4.1.2.5 Lower Geyser Basin

Pocket basin is located on the right (northwest) side of the Firehole river in the western part of the Lower Geyser Basin, about 1.5 km west of the road Madison - Old Faithful. It is, like “Verde Pool” in Norris Geyser Basin (section 4.1.2.3), interpreted as a hydrothermal explosion feature (Muffler et al. 1971). The shallow reservoir temperature underneath Lower Geyser Basin of 200-215°C is significantly cooler than in the Norris-Mammoth corridor. The geothermal waters of Lower Geyer Basin are typically alkaline, low in chloride, and high in bicarbonate as a result from dilution of the deep Cl-rich hydrothermal solutions with cold groundwater before subsurface boiling (Ball et al. 1998a, 1998b, 2001; Kennedy et al. 1987; Stauffer and Thompson 1984; Truesdell and Thompson 1982). In one hot spring within Pocket Basin, Azure Spring ( Figure 41), significantly increased concentrations of S2O3

2- (55 µmol/L) were found besides concentrations of 10.6 µmol/L H2S. The high thiosulfate concentrations might be explained either by oxidation of H2S from thermal fluids that mix with oxygenated ground waters before they reach the surface or by hydrolysis of elemental sulfur and interaction with sulfate-bearing minerals that could stem from old solfatara deposits underlying the explosion crater (Xu et al. 1998). Total As concentrations are 1-2 mg/L (Ball et al. 1998a, 1998b, 2001; Kennedy et al. 1987; Stauffer and Thompson 1984).

0 90 180 270 360 450 m N

Geyser Springs group

Artist PaintPots

road Norris - Madison Junction

8

1

2

3 4

6

5

7

Paintpot Hill

Inset

1st level

3rd level

4th level

Geyser Creek

Geyser Creek

A O

2nd level



Figure 41 Pocket Basin at the Firehole river (Lower Geyser Basin); num-bers 1-3 indicate sam-pling sites YNPLG01 - YNPGG03; 2 = “Bath” hot spring, 3= Azure Spring, O = Ojo Cali-ente hot spring, M = mud pot, C = “Cavern” hot spring

4.2 Methodology

Site information, pictures, devices used, and data obtained and corrected, from the sampling of Yellowstone geothermal areas described in the following are presented in an Access database on the enclosed CD (detailed information and instructions see Appendix 9).

4.2.1 Field methods

4.2.1.1 Reconnaissance trip 2002

The field work at Yellowstone was conducted over three sampling campaigns. The first campaign from June, 26th, to July, 2nd, 2002 was a reconnaissance study where sampling focused on dissolved and volatile As species only. Ragged Hills (9 sites), Hazle Lake (3 sites), and Frying Pan Spring (2 site) were selected as the first study areas. Ragged Hills was chosen because it is the most active geothermal area in the Park with a wide range of hydrogeochemical conditions, and one of the highest known total As concentrations (section 4.1.2.3). The high Hg methylation capacity previ-ously detected by USGS researchers (section 4.1.2.2; Table 10) made the Hazle Lake area an inter-

0 55 110 165 220 275 m N


Pocket Basin

Firehole river

Firehole river

1

2

3

parking lot Fountain Flat Drive road Madison - Old Faithful

O

Fairy Creek

M

C

Methodology 101

esting study site for As methylation and volatilization too. Frying Pan Spring was selected because of its location in the acid alteration area with lower As and high H2S and S2O3

2- concentrations (sec-tion 4.1.2.1). In addition to the three study areas, one mud pot and one hot spring 1.8 km northwest of the parking lot for Norris Geyser Basin were sampled in cooperation with the research group of Dr. Kirk Nordstrom, USGS, Boulder.

Speciation of dissolved As was achieved by the ion exchange technique of Le et al. (2000c), de-scribed in section 2.7.2. The ion exchangers were custom-made 5 mL disposable syringes, filled with 1500 mg of cation exchanger material (resin based DOWEX 50WX8 analytical grade, SERVA Electrophoresis, exchange capacity 1.9-2.0 meq/L), and anion exchanger material (silica based DOWEX 1X8 analytical grade, SERVA Electrophoresis, exchange capacity 1.1-1.2 meq/L). Silanized glass wool stoppers were used at the lower and upper end of the cartridges to prevent exchanger material loss. Le et al. (2003c) detected no breakthrough of As(V) when applying 250 mL of a 2 mg/L As(V) solution on 500 mg of a similar silica-based anion exchanger, i.e. the As sorption capacity was more than 500 µg. For the concentration of negatively charged As species expected (< 0.02 µg in 20 mL) this capacity is sufficient even if competing reactions with other ions are considered. Before sampling, the ion exchangers were conditioned with 6 mL 50% metha-nol, then rinsed with 6 mL distilled H2O and kept wet until use. Arsenic separation was performed on site immediately after taking a 20 mL water sample with a disposable syringe. The sample was filtered through a 200 nm cellulose acetate filter (Membrex) before passing the cation and anion exchanger. Flow velocity was approximately 0.5-1 mL/min. The filtrate was collected in a 50 mL PE bottle. All bottles used in the field were new, cleaned with HCl and rinsed with deionized water in the laboratory before use. The ion exchange cartridges were separated, wrapped in aluminum foil and stored in the refrigerator together with the PE bottles until elution at the most 4 weeks later. Own experiments and those of Le et al. (2000c) confirm that As on the exchangers can be quantita-tively recovered after that time.

Volatile metallics were sampled with the optimized screening method described in section 3.2.2. For trapping, PTFE bottles containing 100 mL of a 1:100 diluted NaOCl solution were used. Ra-schig rings were added to increase reaction times. Sampling was performed both inside the hot springs via the PTFE collector cells described in section 3.1, and above the hot springs via inverted polyethylene (PE) household boxes as has been described for other projects before (section 2.7.5.1). For sampling volatile metallics dissolved in water, the collector cell inside the hot spring was attached to a 400 mL PTFE bottle via a PTFE capillary that directed the gases inside to the bottom of the bottle. A pump was connected by a valve to the second connection of the PTFE bot-tle. Approximately -350 mbar vacuum was applied to the whole system and the valve was closed, so that the vacuum could only be replaced by gas flow through the membrane of the collector cell. The vacuum container originally designed ( Figure 17) had to be dismissed due to transport restric-tions. The upper half of the 400 mL oxidizing bottle that was only half-filled with oxidizing solu-tion, was used as vacuum space instead. The equipment was left in the field for 1-4 days for cumu-lative sampling. Whenever the vacuum was equalized (i.e. when a gas volume of approximately


200 mL ⋅ 0.35 = 70 mL had replaced the vacuum), vacuum was applied again. After sampling, the oxidizing bottle was detached from the collector cell, and the Raschig rings separated from the oxidizing solution by a sieve. The oxidizing solution was collected in 50 mL PE bottles, and stored in the refrigerator at 6°C until analysis.

The collector cell method turned out to be only successful in a few places with low temperatures, such as Hazle Lake. Especially in the hot springs at Ragged Hills, high temperatures deformed the teflon, widened the tight connections, and finally led to water breakthrough. For further sampling, the collector cells were dismissed and only gas samples were taken. The inverted polyethylene (PE) boxes used as gas hoods were approximately 30 cm long, 20 cm wide, and 10 cm high. Hot springs were only sampled where the water was shallow enough (few centimeters of water depth) to place the inverted PE boxes on the ground fixed by some stones. Surrounding water provided a gas-tight seal, so that gas could only leave the inverted PE box through the teflon tube outlet on top of the box connected to the oxidizing bottle. The outlet of the oxidizing bottle was left open. A T-shaped, open valve was used to protect the oxidizing solution from contamination by dust and rain (see also Figure 42). No active pumping was applied to this gas sampling setup and no gas flow was detected at the outlet of any oxidizing unit. Obviously, the pressure built up underneath the gas hood was insufficient to overcome the backpressure of the approximately 20 cm high column of oxidizing solution. Thus, trapping of volatiles was based only on equilibration of concentration gradients between the gas phase under the gas hood, and the oxidizing solution by diffusion. No species-selective sampling was done.

During the reconnaissance trip, volatile As was detected in all study areas. It was the first discovery of volatile As in the Park.

4.2.1.2 Sampling trips 2003

Based on the positive results from 2002, two hydrogeochemical sampling campaigns were carried out from June 5 to July 16 and from September 2 to October 12, 2003. In the water samples, tem-perature, pH, conductivity, redox potential, dissolved oxygen, Fe(II)/Fe(tot), S(-II), SiO2, As spe-cies (As(III) / As(V) / MMAA / DMAA), major anions, and cations, trace elements, TIC, and TOC were determined. Gas samples were analyzed for volatile HCl, H2S, and volatile metallics as the sum of all species oxidized in NaOCl, and species-selective for As. Besides re-sampling Ragged Hills and Hazle Lake, studies were extended from Frying Pan Spring to the Nymph Lake area, in-cluding the recently appeared fumaroles on the north side of the lake, and to two new study areas. Geyser Spring Group in the Gibbon Geyser Basin, located on the south side of Norris dome, shows a wide variety of hydrogeochemical conditions and As concentrations from 0.1-2.4 mg/L like Nor-ris Geyser Basin. However, it does not seem as dynamic as Ragged Hills at the moment, and no seasonal variations in geothermal activity have been detected there yet (section 4.1.2.4). Pocket Basin in Lower Geyser Basin was the only study area that was located within the Yellowstone cal-dera and showed high As concentrations in alkaline waters (section 4.1.2.5). A total of 6 sites in the Nymph Lake / Frying Pan Spring area, 3 sites at Hazle Lake, 28 sites at Ragged Hills, 8 sites at

Methodology 103

Geyser Spring Group, and 3 sites in the Pocket Basin were sampled one to three times during the two campaigns in 2003. Appendix 10 gives an overview of the parameters investigated at each site. Details on sample preparation, devices and methods used, as well as necessary corrections done are explained in Appendix 9 and together with the respective data in the digital database.

Sampling sites were located by a handheld GPS (Garmin-12). In the Ragged Hills area a differen-tial GPS (Pathfinder Pro XR Trimble) with a higher precision was used during a GIS-project on high resolution mapping of the geothermal features (Becker 2004). Water temperature, conductiv-ity, pH, redox potential and dissolved oxygen were determined on site with a WTW (Wissen-schaftlich-Technische Werkstätten GmbH Weilheim) MultiLine P4 device by immersing probes directly into the source. Water temperature was measured both with the SenTix 97/T pH electrode and the TetraCon 325 conductivity cell. Generally, mean values of the two readings were taken. Both the redox Pt 4805/S7 probe and the CellOx 325 probe quit working at temperatures >50°C and measurements were dismissed for most samples with higher temperatures. Occasionally redox potential and dissolved oxygen were measured in a container after cooling down to <50°C. How-ever, because no flow-through cell and pump were available, these values are questionable. The redox potential measured against a Ag/AgCl electrode was corrected for standard hydrogen poten-tial and temperature according to the following equation based on information provided by the manufacturer (WTW): EH(corrected) = EH(measured) + (-0.7443 ⋅ temperature [°C] + 224.98).

Total Fe, Fe(II), SiO2, and S(-II) were determined on site by field photometry. A 20 mL disposable PE syringe rinsed 5 times with sample water was used to withdraw 10-25 mL sample directly from the sampling site. The sample was immediately filtered through a 200 nm cellulose acetate filter (Membrex) and analyzed with a HACH photometer DR/890 applying the FerroVer, 1,10-phenanthroline, silicomolybdate, and methylene blue method for total Fe, Fe(II), SiO2, and S(-II) respectively. The Fe(III) concentration was afterwards calculated as the difference between total Fe and Fe(II).

Commercial cation and anion exchangers were used for dissolved As speciation during the 2003 sampling campaigns. The strong resin-based cation exchanger with a particle size of 45-150 µm, and an exchange capacity of 1.9 meq/mL was obtained from Alltech, the silica based anion ex-changer with a particle size of 50 µm, and a total capacity of 0.14 meq/g from Supelco. The 4 mL cartridges contained 500 mg exchanger material between 1 mm thick, porous disks. The disks proved advantageous compared to the glass wool plugs in the homemade exchangers used in 2002. They provided more uniform packing heights and caused no distortion of the exchanger material when applying the sample. As for on-site photometry, 20 mL sample were withdrawn from the sampling site by a disposable PE syringe and filtered through a 200 nm cellulose acetate filter (Membrex) before passing the ion exchangers with a flow rate of approximately 0.5-1 mL/min. The As(III) was collected in a 50 mL PE bottle. The ion-exchange cartridges were separated, wrapped in aluminum foil and stored in the refrigerator at 6°C together with the PE bottles until elution in the laboratory at most 4 weeks later.


All bottles used for sampling were soaked with 10%HNO3 and rinsed with deionized water in the laboratory prior to sampling. Samples were taken unfiltered for IC (200 mL PE bottle), unfiltered for TOC/TIC (100 mL glass bottle), and filtered through a 200 nm cellulose acetate filter (Mem-brex) and stabilized with 1 mL HNO3 suprapur (50 mL PE bottle) for ICP-MS, ICP-AES, and HG-AAS and GF-AAS. The bottles for IC and TOC/TIC were rinsed 5 times with unfiltered sample water, the bottles for ICP-MS, ICP-AES, HG-AAS, and GF-AAS with filtered sample water before filling them completely. Filtration was done immediately after withdrawing the sample from the sampling site with a disposable 20 mL PE syringe. At selected sites, Hg samples were taken. For inorganic Hg the water sample was filtered through a 200 nm filter into a glass bottle and 5 mL HNO3 were added directly in the field. The methylated Hg samples were filtered through a 200 nm filter into a PE bottle, and 10 mL 50% HCl were added to 500 mL sample upon return from the field the same day.

After the experiences from the 2002 reconnaissance trip, volatile As was sampled with gas hoods only, and trapped in 100 mL of 1:100 diluted NaOCl. The gas hoods used were inverted PE boxes of different sizes (33 x 24 x 8 cm, 30 x 20 x 10 cm, diameter 30 cm x 13 cm). The equipment was left outside for 52 to 262 hours for cumulative sampling. Oxidizing solutions were then refilled in 50 mL PE bottles, and stored in the refrigerator at 6°C until analysis. Furthermore, PTFE bottles with 100 mL of a 0.1 N NaOH solution were used to oxidize and dissolve volatile HCl. The NaOH concentration was lower than cited in Giggenbach (1975) or Fahlquist and Janik (1992) who use 4 N NaOH. However, those experiments were done by purging a bubbling gas stream through the trapping solution, resulting in short reaction times. As mentioned before, the gas sampling de-scribed here was based on diffusion without apparent gas flow and with longer reaction times. Lower NaOH concentrations are advantageous since they show less matrix effects during analysis. After cumulative sampling, the NaOH solutions were also stored in 50 mL PE bottles in the refrig-erator at 6°C until analysis.

Mainly during the third sampling campaign, different setups of gas hoods and oxidizing bottles were tested to determine: • the decrease in concentration of volatile metallics above a sampling site by comparing the

concentrations directly above the sampling site taken from within a gas hood to concentra-tions in ambient air approximately 25 cm above the sampling site taken without any funnel device (setup 1 and 3, Figure 42)

• the effectiveness of gas diffusion and dissolution in setups with the tube immersed in the oxi-dizing solution compared to setups where the tube is left in the headspace above (setup 1 and 2, Figure 42)

• the potential of contamination when preparing and re-filling the oxidizing solution in the field (field blank; setup 4, Figure 42)

Three samples of setup 1 (YNPNL02-1G; YNPNL03-3G; YNPRH11-4G), one sample of setup 2 (YNPRH23-4G), and one sample of setup 3 (YNPRH11-4G) were not considered for further inter-

Methodology 105

GH GH

1 2 3 4

NaOCl / NaOH

NaOCl / NaOH

NaOCl / NaOH

NaOCl

pretation, since the rubber sealing of their oxidizing bottle was displaced and contamination of the sample could not be excluded.

Figure 42 Different setups for diffusive sam-pling (setup 1-3); setup 4 was used to check acci-dental contamination effects

1 = Standard setup: tube from the gas hood (GH) is immersed in the oxidizing solution, outlet (on the right) of the PTFE

bottle is protected with an open T-shaped valve from contamination from outside [samples: 51 NaOCl, 37 NaOH] 2 = Headspace: PTFE bottle is connected to a gas hood (GH), but tube ends in the headspace above the oxidizing solu-

tion; outlet: open T-valve [samples: 4 NaOCl, 2 NaOH] 3 = no gas hood, oxidizing bottle left with 2 open T-valves adjacent to the sampling site [samples: 19 NaOCl, 10 NaOH] 4 = no gas hood, oxidizing bottle left with 2 closed T-valves adjacent to the sampling site (field blank) [sample: 1

NaOCl]

In addition, flat PE boxes (10 x 10 x 3 cm and ∅ 7 cm x 3cm) filled with 100 mL 1:100 diluted NaOCl were placed adjacent to sampling sites at Hazle Lake (YNPHL01, 03), Frying Pan Spring (YNPNL02, 03, 04), and Ragged Hills (YNPRH01, YNPRH10). A PE lid was fixed on top of the “diffusion bowls” (DB) to prevent dust and particles from entering the oxidizing solution. Tiny rubber stoppers taped on the rim between PE lid and bowl left a 1 mm slit open for gas diffusion. The DBs were thought to be advantageous compared to the oxidizing bottles due to their larger reaction surface, however, significant evaporation losses were encountered for the oxidizing solu-tion inside and samples could only be obtained from 21 of 43 setups (for details see digital data-base, and Appendix 9).

At four sites (YNPGG04, YNPHL01, YNPRH01, and YNPRH06), gas samples were obtained through gas hoods by active gas pumping with a Dräger hand pump (volume per stroke 0.1 L). Total volumes of 5 to 20 L were sampled in a period of 1 to 4.5 hours. An additional PTFE bottle was used as H2O trap before the oxidizing solution. The amount of condensed water from the hot gases was 150 to 300 mL.

For species-selective sampling of volatile As, solid sorbents and SPME fibers as described in sec-tion 3.2.3 were used. The solid sorbents were custom-made single bed tubes of 250 mg Carboxen 564 mounted behind commercial SKC Tenax TA double layer tubes ( Appendix 4). The Tenax TA tubes served as a pre-separation step to trap larger molecules in the gas samples, so that the Car-boxen was supposed to be exclusively loaded with substance of higher volatility especially the target element As. Because no GC-MS with desorption port was available in the vicinity of the


study area, the tubes were wrapped in aluminum foil, and stored in a refrigerator at 6°C until they were sent to Germany by express mail. The analyses were performed 9 to 15 days later.

Because of their different performance in trapping individual As species, three different SPME fiber types (PDMS 100, PDMS-CAR, PDMS-CAR-DVB) were used. For the field application, a 2 mL plastic syringe was used as fiber holder ( Figure 43). The fiber is protected inside the sy-ringe´s needle. The syringe plunger is used to press the spring together and expose the fiber inside the needle. The third way of the T-shaped valve works as an open gas outlet. A PE tube attached over the needle to the T-shaped valve connects the fiber holder to the gas hood (GH)

Figure 43 Custom-made SPME fiber holder for field application (fiber is protected inside the syringe needle; syringe plunger is used to expose the fiber; third way of T-shaped valve functions as gas outlet)

4.2.2 Laboratory

Ion exchangers for As speciation were eluted in the laboratory at the Technische Universität Ber-gakademie Freiberg. A volume of 10 mL of 60 mM acetic acid was used to desorb MMAA from the anion exchanger. Using 10 mL provided an enrichment factor of 2 compared to the 20 mL sam-ple applied to the exchangers in the field resulting in a higher accuracy and lower detection limits. The As(V) left on the exchanger was eluted with 1N HCl applied in 10 mL steps until As concen-trations in the eluent dropped below detection limit. Similarly, DMAA was desorbed from the cation exchanger in 10 mL steps with 1 N HCl. The acetic acid, HCl eluents, and the water sample with the As(III) fraction were analyzed by a Zeiss AAS EA4 spectrometer using graphite furnace with platform technique. Calibration was done with matrix-matched standards. Total As was ana-lyzed separately from a filtered and HNO3-acidified sample. Detection limits, quantification limits, and upper calibration ranges for this and all following determinations are given in the digital data-base as well as results from double and triple determinations and verification of reference material for quality control.

The TOC and TIC were determined in Freiberg with a LiquiTOC (Elementar Analysensysteme GmbH). Concentrations of Cl, S(VI), F, Br, and N(V) were determined by IC (Dionex model 2010i ion chromatograph with AG4A guard and AS4A separator columns and an anion micromembrane suppressor-II column; eluent: 0.018 M NaHCO3 + 0.017 M Na2CO3) at the U.S. Geological Survey laboratories in Boulder, Colorado. Some samples were re-analyzed at the Technische Universität Bergakademie Freiberg with a Merck Hitachi D 6000 (conductivity detector; eluent: anions: 415 mg/L pthalic acid + 278 mg/L tri(hydroxymethyl)-aminomethane; cations: 167 mg/L pyridine-2,6,dicarbonicacid + 750 mg/L 2,3 dihydroxyamberacid).

gas flow from GHplunger

2 mL syringe T-shaped valve

gas outlet

needle

spring to expose the fiber

fiber tip with reactive coating

fiber metal shaft

PE

Methodology 107

The ICP-AES (Leeman Labs Direct Reading Echelle, dual view, sequential, multi-element, induc-tively coupled plasma spectrometer, Hildebrand grid nebulizer, glass Scott spray chamber) at the USGS laboratory was used to measure As, Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Si, Se, Sr, V, and Zn. Information about scan mode (axial or radial), and wave-length for each element as well as about control material is given in the digital database. Selected samples were sent to Activation Laboratories, Canada (accredited by ISO 17025), for ICP-MS analysis on Li, Be, Na, Mg, Al, Si, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr, Y, Zr, Nb, Mo, Ru, Pd, Ag, Cd, In, Sn, Sb, Te, I, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Ta, W, Re, Os, Pt, Au, Hg, Tl, Pb, Bi, Th, and U (code 7, natural waters). Mercury was analyzed by cold-vapor atomic-fluorescence spectrophotometry (CVAFS) with a PS Analytical, model Galahad, at the USGS in Boulder, Colorado (Dr. Howard E. Taylor).

The NaOCl gas samples were analyzed by IC in Freiberg for dissolved S(VI). The ICP-AES in Boulder was used to analyze the NaOCl samples for the elements mentioned before. The HG-AAS in Boulder (Perkin-Elmer Analyst 300 with a FIAS-100 flow injection analysis system hydride generator) was applied to determine As concentrations. Pre-reduction was initially done with 1.2 mL 10% HCl, and 0.5 mL KI in ascorbic acid on 0.5 mL sample (diluted to a total of 5 mL with deionized H2O). Because the June / July samples showed a greater variance between multiple de-termination, reducing agents were doubled for the September / October samples and the samples were allowed to pre-reduce for 1 hour before analysis to achieve a reproducible, complete reduc-tion. Selected gas samples were sent for ICP-MS analysis to Actlabs. For NaOCl samples from diffusion bowls only the main target element As (HG-AAS) was determined due to the low amount of sample volume. Chloride in the NaOH samples was titrated in Freiberg by the mercury thiocy-anate method. In adaptation of the standard procedure, the 0.1 N NaOH samples were first neutral-ized to pH 7 with 40 µL HClO4 per 5 mL sample. The samples were diluted with 20 mL H2O be-fore the mercury thiocyanate reaction.

Solid sorbents were analyzed in Freiberg with the Bruker mobile GC-MS EM 640 and the opti-mized conditions listed in Table 8. The SPME fibers were analyzed at the United States Federal Center in Denver, Colorado with a Varian 3400 in June/July 2003, and a Varian CP 3800 in Octo-ber 2003. Both times the column was a HP5ms (L=30 m, ID=0.25 mm, fd=0.25 µm). Optimized conditions for SPME analysis can be found in Table 9. Additionally, in September and October 2003 analyses were conducted with a HP 5890 GC-MS with a DB 5 column (L=30m, ID=0.32mm, fd=0.25um) at the Department of Chemistry, Montana State University, Bozeman.

4.2.3 Data processing

4.2.3.1 Water data check

All water analyses were checked for accuracy by calculating the charge imbalance with the hydro-

geochemical modeling program PHREEQC 2.8.03 (Parkhurst and Appelo 1999) and the database


WATEQ4F (V.2.4 with revised data from Nordstrom and Archer 2003) using the following equa-

tion:

[meq/L]) Σanions q/L]cations[me([meq/L]) Σanions - q/L]cations[me(*100errorpercent

+ΣΣ

=

Negative analytical errors indicate a deficiency of cations or a surplus of anions, positive values a surplus of cations or a deficiency of anions. If errors compensate each other, e.g., in an overestima-tion of both a cation and an anion, they remain undetected with this check. An additional independ-ent accuracy check was therefore applied using the difference between measured and calculated conductivity as proposed by Rossum (1975) and utilized in WATEQ4F (Ball and Nordstrom 1991; Table 11).

Table 11 Statements from accuracy checks of water analyses based on the calculation of the charge balance and a comparison of measured and calculated conductivity

negative charge balance positive charge balance

measured conductivity > calculated conductivity deficiency of cations deficiency of anions

measured conductivity < calculated conductivity surplus of anions surplus of cations The conductivity is calculated from the equivalent conductance of the individual cation and anions

in infinitely dilute solutions and their valence states according to the Kohlrausch Law:

G = Λ0⋅C - (K1⋅Λ 0 + K2) ⋅ C1.5

G = conductivity of the solution C = concentration of the solution Λ0 = equivalent conductance

K1 and K2 are constants associated with the relaxation of the ion-cloud effect and the electropho-retic effect relative to ion mobility. They are calculated from temperature, viscosity, and dielectric constant of water, and depend on the equivalent conductance and valence of each ion. For 25°C, a viscosity of 0.008949 poise, and a dielectric constant of 78.55 conductivity is calculated as:

[ ]1.5-

--

00 )CZ(Z0.668Q1

2Q)Z(Z115.2

ZZΛGG G +⋅

+

+⋅

+⋅⋅⋅

−+= ++

+−+

with Λ0⋅C = G0+ + G0- and −− ⋅= ζΣc G -0 and +++ ⋅= ζΣc G0 ; ζ = equivalent conductance

−−

−−⋅Σ⋅Σ

=zczc 2

- Z ++

+++ ⋅Σ

⋅Σ=

zczc 2

Z

−

=ΣcG

λ 0-

++ =

ΣcG

λ 0 Λ0 = λ+ + λ-

)ZλZ(λ)Z(ZZZΛ Q

----0

+++

++⋅+

⋅⋅=

Extrapolated values for equivalent conductance of H+, OH-, HCO3-, CO3

2-, HSO4-, SO4

2-, NO3-, Cl-,

F-, Br-, Na+, K+, Li+, Ca2+, Mg2+, Ba2+, Fe2+, Fe3+, Al3+, and Sr2+ in the standard state of infinitely

Methodology 109

dilute solutions at 25°C were taken from Rossum (1975), completed by those listed in Coury (1999). Concentrations of each species were determined by speciation modeling with PHREEQC 2.8.03 (database WATEQ4F). For modeling, values below detection limit were replaced by 0.3 ⋅ detection limit, missing values were dismissed. Concentrations of H+ and OH- were always consid-ered because of the wide range of pH from 1.8 to 8.6 covered by the individual samples. Within the pH range of 5 to 8, neither H+ nor OH- will contribute significantly to conductivity. From a total of 60 water samples, 34 samples showed analytical errors less than 2%, 15 samples 2-5%, and 11 samples >5%. The maximum analytical error was 8.3% (detailed results in the digital database).

Based on the results from the data check, all raw data were revised again, some re-analyzed. From double determinations in different dilutions or on different equipment the value with the better fit for the analytical error and a lower range of analysis of measured to calculated conductivity was selected. One of the parameters that was especially carefully checked was pH. In acid waters H+ is the predominant cation, and slight changes in the pH can already alter the ion balance significantly. Field pH was generally considered to be the most accurate. In addition, the pH of all samples was measured again in the laboratory. For samples, where the analytical error was more than 2% with the field pH, the laboratory pH was used if it resulted in a better charge balance. If speciation mod-eling showed a better charge balance with pH adaptations of ± 0.1 pH units, the modeled pH was accepted. Results of the revised pH and adaptations of other raw data are documented and justified in the digital database.

Values from As speciation were checked in comparison with total As determination by GF-AAS in Freiberg and ICP-AES in Boulder. The range of analysis (ROA) from GF-AAS and ICP-AES is 4.4% on average. If concentrations below 100 µg/L are not considered, ROA is 2.5% on average, with a maximum of 7% and a minimum of 0.06%. For concentrations less than 100 µg/L, GF-AAS is the more accurate method. The range of analysis of As calculated from the sum of the four spe-cies compared to As from ICP-AES (>100 µg As/L) was 3.1% on average with a maximum of 9%, compared to As from GF-AAS 2.6% on average with a maximum of 11%. Only in one case (YNPNL01), the As concentration determined by speciation was obviously too low (107 µg/L compared to 160 µg/L by GF-AAS, and 164 µg/L by ICP-AES). In all other cases, the values were accepted as being in good agreement, considering that analytical errors of 4 single determinations influence the As concentration calculated from the sum of species.

The As speciation technique was originally developed for waters with near neutral to slightly acid pH only (section 2.7.2). It is influenced by the extreme pH some samples have. Below pH 4, part of the As(V) fraction occurs as non-charged complex ( Figure 4). This non-charged complex passes the ion-exchanger and is thus (erroneously) determined in the As(III) fraction. For the lowest pH measured (1.94), the non-charged complex presents 60% of the As(V) fraction. Thus, the As(V) concentration might be too low by a factor of 2.5. Since the samples with low pH show As(V) con-centrations in the range of a few µg/L only, compared to several 100 µg/L in the As(III) fraction, the error is minimal. For MMAA, the non-charged complex becomes predominant already at a pH below 3.6 ( Figure 8). It should thus be expected that a significant amount of MMAA is not sorbed


on the anion exchanger. However, eluted MMAA fractions of 2-19% of all As species with an av-erage of 8% are high compared to concentrations from literature (section 2.5.2.1). Assuming even higher concentrations seems implausible. These reflections could suggest that MMAA was com-pletely bound to the exchanger. Considering that MMAA predominates as non-charged complex at pH below 3.6, an effective binding would require an additional interaction of MMAA with the matrix of the anion exchanger, not only sorption on the functional groups. These hydrophobic in-teractions are known for DMAA (section 2.5.2.1), but seem less likely for a compound with only one methyl group. Another explanation for the high As concentrations in the MMAA fraction is a co-elution of MMAA and As(V) compounds that are only weakly bound at low pH. Especially for sites where DMAA concentrations are low, the “MMAA” fraction is not much higher than the as-sumed percentage of non-charged As(V). To be consistent, the terms As(III), As(V), MMAA, and DMAA fraction were used throughout the report according to the method described in section 2.7.2. For interpretation, it has to be considered that only DMAA is positively identified because its sorption is independent of charge. The “As(III) fraction” may contain some non-charged As(V) and MMAA compounds, and thus is more correctly referred to as the fraction of non-charged As spe-cies at the individual pH. The “MMAA” fraction is the fraction of weakly bound, negatively charged As. It might contain some As(V) compounds. The “As(V)” fraction consists of the strongly bound, negatively charged As species that presents only a minor part at low pH.

The concentrations of Fe species were compared as determined by on-site photometry and ICP-AES. The overall ROA is 5.8% for Fe concentrations > 0.2 mg/L. On-site determinations were taken for all further modeling, because they contain species information on Fe(II)/Fe(III). Compari-son between SiO2 from on-site photometry and ICP-AES showed a ROA of 8.2% with a systematic error of photometric values being lower than ICP-AES values in 55 of 60 cases. Interference from S(-II) or Fe might be responsible for this reduction. Thus, ICP-AES values were taken for further modeling.

4.2.3.2 Gas concentration conversion

Volatile compounds were trapped in NaOCl and determined as their respective dissolved, oxidized species by GF-AAS or ICP-AES as described in the sections 4.2.1 and 4.2.2. For a quantitative interpretation of volatile compounds degassed from each sampling site, the dissolved concentra-tions in µg/L had to be converted into gaseous concentrations in µg/m3. Different approaches had to be used for the sites of passive gas sampling with different setups of gas hoods – oxidizing bot-tles ( Figure 42), diffusion bowls, and active gas sampling with pumping.

Passive gas sampling Passive gas sampling is mainly based on diffusion. The net diffusion rate for a volatile compound depends on its diffusion coefficient, the concentration gradient and the diffusion length according to Fick´s first law:

∆x∆cD J ⋅−=

Methodology 111

J net diffusion rate = mass per time and diffusion area: Atm J⋅

=

m mass [µg] = concentration of dissolved, oxidized compound [µg/L] ⋅ volume of trap-ping solution [100 mL]; analytical precision of mass is about ± 5% for Cd, Cr, Co, Pb, Sb, and Se, and ± 1-2% for all other elements, including As

t exposure time (for the 2003 sampling between 52 to 262 hours) [s] A diffusion area; flow cross-section of tube connecting sampling site and trapping solu-

tion; tube diameter = 2 mm (r = 0.1 cm); A = r2 ⋅π = 0.0314 cm2 [cm2] D diffusion coefficient [cm2/s] ∆c concentration gradient between sampling site and trapping solution [µg/cm3] ∆x diffusion length; length of the tube connecting sampling site and trapping solution [cm]

Assuming a 100 % effective dissolution and oxidation, the concentration of the volatile compound in the trapping solution and at the gas-solution interface is 0 during the whole sampling time. Thus, the concentration gradient ∆c equals the concentration c at the sampling site in µg/cm3:

DAt∆xm c⋅⋅

⋅=

The diffusion coefficient varies for individual gases and with temperature. If no values are avail-able from literature, as in the case of all volatile As species, the Fuller, Schettler and Giddings (FSG) equation can be used for a semi-empirical estimation of the diffusion coefficient based on molecular volumes (Fuller et al. 1966; Tucker and Nelken 1990):

2

0.51.75

)V(V P

MT0.001 D

1/31/3BA

r

−⋅

⋅=

D diffusion coefficient [cm2/s] T absolute temperature [K] above sampling site, estimated as the average of water temperature

at the sampling site and average ambient air temperature (15°C); for the small fumaroles, where no water temperature was measured, an average temperature of 45°C was assumed

Mr BA

BAr MM

MM M

⋅+

=

MA = molecular weight of air (approximately 28.97 g/mol) MB = molecular weight of compound of interest [g/mol] P pressure [bar] VA molar volume of air (approximately 20.1 cm3/mol) VB molar volume of the compound of interest [cm3/mol] = molecular weight [g/mol]: vapor den-

sity [g/cm3]

Typical values for diffusion coefficients of inorganic and organic gases in air are between 0.1 to 0.2 cm2/s. To achieve a higher accuracy, information about the exact composition of the gas including the percentage of each species for each element (e.g., AsH3, MMA, DMA, TMA, etc.) would be needed. A sensitivity analysis was done for volatile As, assuming an uncertainty of ± 10°C for the


temperature estimation above the sampling site, and the two extreme cases that all volatile As is present as the lightest volatile As species AsH3 (MB = 77.945 g/mol; VB = 77.945 g/mol : 2.70 g/cm3 [Sax 1986] = 28.87 cm3/mol) or as the heaviest volatile As species AsCl3 (MB = 181.28 g/mol; VB = 181.28 g/mol : 6.25 g/cm3 [Clayton and Clayton 1981] = 29 cm3/mol). Calculated dif-fusion coefficients ranged from 0.115 cm2/s to 0.172 cm2/s with a mean of 0.147 ± 0.015 cm2/s. This maximum deviation of 10% shows that neither temperature nor species are very sensitive pa-rameters. Thus, a mean value from the four calculations was taken for each sampling site. The dif-fusion coefficients calculated for As were also accepted as an approximation for the other volatile compounds determined by ICP-AES. For all volatile Cl determined from the NaOH trapping solu-tion, volatile HCl was assumed to be the only gaseous species (MB = 36.461 g/mol; VB = 36.461 g/mol : 1.268 g/cm3 (Merck 1983) = 28.75 cm3/mol). Considering a temperature inaccuracy of ± 10°C, diffusion coefficients were between 0.145 and 0.196 cm2/s with a mean of 0.176 ± 0.010 cm2/s. A mean value from the calculation with the two temperature extremes was taken for each sampling site. Calculations are documented in the digital database ( Appendix 9).

For the different sampling geometries, i.e. the different gas hood – oxidizing bottle setups, and diffusion bowls, calculations were performed as explained in the following. Because all oxidizing bottles had an inlet and an outlet on top for the gases to enter, the total net diffusion rate is the sum of the diffusion through inlet and outlet:

)∆x∆c

D(∆x∆c

D J2

22

1

11 ⋅−+⋅−= where D1 = D2

Calculation varies for the 4 different setups chosen ( Figure 42): • Setup 4: An oxidizing bottle was left at the sampling site with inlet and outlet being closed by

valves to check on contamination of the samples in the field during preparation and filling of the oxidizing bottles or diffusion through caps or tube connections. This field blank showed values comparable to the laboratory blanks of a 1:100 diluted NaOCl solution, indicating that there was no contamination, and no diffusion with the valves closed.

• Setup 3: The oxidizing bottle in this setup was not connected to a gas hood and both valves were left open. This setup gives information about volatile gas concentrations in approxi-mately 25 cm above ground. It was assumed that a stable layer of air with increased gas con-centrations may form under calm wind conditions over the geothermal features. Diffusion coefficients, diffusion lengths and concentration gradients for this setup are identical for inlet and outlet. The net diffusion rate is:

)∆x∆cD(2)

∆x∆c

D(∆x∆c

D J2

2

1

1 ⋅−⋅=⋅−+⋅−= and DAt2∆xm c

⋅⋅⋅⋅

=

with ∆x = 5 ± 2 cm; A = 0.0314 cm2, and D as discussed above

• Setup 1 and 2: While setup 3 only gives information about concentrations in the ambient air about 25 cm above the sampling site, setup 1 and 2 integrate concentrations from ambient air in 25 cm above the sampling site (via the short tube) and concentrations from directly above

Methodology 113

the sampling site (via the long tube connected to the gas hood). The following equation can not be solved due to the two unknown concentrations ∆c1 and ∆c2:

)∆x∆c

D(∆x∆c

D J2

2

1

1 ⋅−+⋅−= and DAtm

∆x∆c

∆x∆c

2

2

1

1⋅⋅

=+

with ∆x1 = 5 ± 2 cm; ∆x2 = 60 ± 20 cm; A = 0.0314 cm2

The concentration gradients ∆c1 and ∆c2 are most likely not identical. The general assump-tion is that concentrations underneath a gas hood should be higher than concentrations 25 cm above the sampling site that are subject to diffusion, turbulent convection (eddy flux), dilu-tion, hydrolysis, and oxidation. However due to the longer diffusion length from gas hood to oxidizing bottle (60 cm) compared to diffusion length from the open air (5 cm) the net diffu-sion rate might be significant from both sides. The problem could have been solved by clos-ing the outlet with a valve, so that only diffusion from the gas hood has to be accounted for. This setup was not realized, because initially it was assumed that some gas flow might occur apart from the net diffusion that would lead to a build-up of pressure inside the oxidizing bottle. In the end, no gas flow was detected at any sampling site during any sampling cam-paign, and there was nothing to change about the setup design anymore at that time.

An approximation that can be made, however, is the calculation of an upper value for both compartments from two end-member models. In the first model it is assumed that there is no concentration gradient between oxidizing solution and gas hood (∆c2 = 0). The total amount of volatile compounds then results from diffusion between ambient air and oxidizing solu-tion (maximum for c1). In the other end-member model there is no concentration gradient be-tween ambient air and oxidizing solution (∆c1 = 0) and all volatile compounds enter through the gas hood (maximum for c2).

DAtm

∆x0

∆x∆c

21

1⋅⋅

=+ DAt∆xm

c 11 ⋅⋅

⋅=

hence DAtm

∆x∆c

∆x0

2

2

1 ⋅⋅=+ and DAt

∆xm c 2

2 ⋅⋅⋅

=

For the 19 sites (13 NaOCl, 6 NaOH samples), where the setups 1 and 3 were run parallel, the con-centration in 25 cm determined by setup 3 was also taken as representative for setup 1. In this case, the net gas flow by the short diffusion path was known (∆c1), and the diffusion by the gas hood (∆c2) could be calculated as:

DAtm

∆x∆c

∆x∆c

2

2

1

1⋅⋅

=+ and )∆x∆c

DAtm(∆x∆c

1

122 −

⋅⋅⋅=

The results gave information about possible enrichment effects underneath the gas hood compared to ambient air 25 cm above the sampling site. At 5 sites (3 NaOCl, 2 NaOH), setup 2 was run to-gether with setup 1 and 3, enabling a direct comparison between gas diffusion and dissolution di-rectly in the oxidizing solution or with the headspace above.


For the diffusion bowls the calculation is

DAt∆xm c⋅⋅

⋅=

A diffusion area is the entrance area for the gases into the bowl defined by the circumference of the bowl and the height of the open slit: 4 ⋅ 10 cm ⋅ 0.1 cm = 4 cm2 for the squared DBs (No. 1-40), 2 ⋅ (7 cm / 2) ⋅ 3.14 ⋅ 0.1 cm = 2.198 cm2 for the round DBs (No. 41-43)

∆x diffusion length is the height from the gas entrance on the top of the bowl to the oxidizing solution; 100 mL oxidizing solution result in 1 cm filling height in the squared DBs, and 2.5 cm in the round DBs; total height 3 cm → 2 cm and 0.5 cm from entrance to oxidizing solu-tion for the squared and the round DBs respectively

Active gas sampling Easier to convert are the gas concentrations from the active gas pumping done at the four sampling sites YNPGG04, YNPHL01, YNPRH01, and YNPRH06 (section 4.2.1.2):

Vm c =

c concentration at sampling site [µg/m3] m mass = concentration of dissolved, oxidized compound in µg/L ⋅ volume of trapping solution

(100 mL) [µg] V sampled gas volume during pumping [m3]

The uncertainty is significantly lower compared to the calculations based on gas diffusion dis-cussed above. It only depends on the accuracy of analytical determination (about ± 5% for Cd, Cr, Co, Pb, Sb, and Se, and ± 1-2% for all other elements including As) and the uncertainty per stroke of the Dräger hand pump (± 5%). The sampled gas volume is well defined and no problems with sampling geometry occurred. Two disadvantages are the longer field working times and the more expensive equipment requirements. Whereas setting up 10 passive diffusive samplers and recover-ing them after a passive sampling time of several days requires a total of about 2 hours, active gas sampling took between 1 to 4.5 hours per site depending on the gas flow from the ground. Dilution might occur during active gas sampling, if the pump rate exceeds the gas flow from the ground and additional ambient air is drawn in. Furthermore, the passive sampling over several days yields an integrated signal, whereas the active sampling for 1-4.5 hours is a “snapshot”. For several sites, passive sampling can be done more or less within the same sampling period whereas active pump-ing has to be done one by one on several consecutive days. The difference might be essential be-cause temporal variations of gas composition are likely to occur, especially for microbially cata-lyzed volatilization of metallics. Pumps with a very low pump rate, supplied by a solar panel left in the field for several days, could also yield an integrative signal. However, this setup is not only more expensive for a screening study with numerous sampling sites but also more sensitive to dis-turbances (breakdown of electrical equipment, wildlife impact) and more difficult to camouflage from tourists and wildlife (National Park Service requirement). Thus passive gas sampling was the general method of choice, completed by active gas sampling at selected sites.

Methodology 115

4.2.3.3 Gas chromatography interpretation

Chromatogram interpretation for solid sorbents was done with the Bruker Data Analysis Software 1.0i. For the interpretation of all SPME data, the public domain mass spectral data reduction com-puter program “Wsearch 32” (Version 01.01.2003) was used. Varian files in the new format (“.sms”; from analysis in Denver) had to be converted to old format (“.ms”) via the “Batchsms” code, Version 1.0 from Varian CSB. Hewlett Packard Pascal file (“.d”; from analysis in Bozeman) could be read directly by the Wsearch 32 software. Peaks were integrated manually. The chroma-togram for each analysis and peak integration and mass spectra of volatile As species are presented in a tabular and graphic form in the digital database. The main objective was a qualitative interpre-tation with the detection of different As species. Quantitative interpretation was dismissed because of the problems discussed in section 3.2.3.2. Thus, no standard calibration was done. Integrated peak areas are given in the digital database for a semi-quantitative comparison. Occasional poor peak resolution from the background or double peaks are shown in the respective figures and con-sidered for final data interpretation.

4.2.3.4 Statistics

Final data interpretation described in section 4.3 is discussed in groups classified by hierarchical cluster analysis via the statistical program SPSS for Windows, version 11.0. The best fit of differ-ent methods (between groups linkage, within groups linkage, nearest and furthest neighbor, cen-troid and median clustering, Ward´s method) with the interval measure squared Euclidian distances was chosen. For the water samples the following parameters were considered: {H+} (10-pH), tem-perature, conductivity, F, Cl, Br, N(V), S(VI), As species (As(III), As(V), MMAA, DMAA), Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Sr, V, Zn, SiO2, S(-II), Fe(II), Fe(III), TIC, and TOC. The volatile elements included for gas sample clustering are As (from GF-AAS), Al, B, Ba, Cu, Fe, K, Li, Si, Sr, Zn, S. All variables were standardized to a range of 0 to 1 and re-scaled before cluster analysis. Values below detection limit were replaced by 0.3 ⋅ detection limit. The significance of each cluster analysis for all variables was checked with the non-parametric Kruskall-Wallis test. The grouping variable was the cluster number according to the different clus-ter methods. The significance of differences between two groups was tested by the Mann-Whitney Test for two independent samples. Bivariate correlation analysis for water and gas samples was done with the nonparametric Spearman correlation coefficient and a two-tailed test of significance.

4.2.3.5 Speciation modeling

Speciation modeling was done with PHREEQC 2.8.03 (Parkhurst and Appelo 1999) using the da-tabase WATEQ4F (V.2.4 with revised data from Nordstrom and Archer 2003). Parameters and species considered were pH, Temp, pe, F, Cl, Br, N(V), S(VI), As(III), As(V), Al, B, Ba, Ca, Cd, Cu, K, Li, Mg, Mn, Na, Ni, Pb, Se, Sr, Zn, Si, S(-II), Fe(II), Fe(III), and C(IV). No thermodynamic data could be found for the organic As species MMAA and DMAA. For modeling they were added according to their most probable redox state +5 to inorganic As(V) to get a better adaptation for the


As redox equilibrium. The concentration of thus-modeled As(V) complexes was referred to the actual concentration of inorganic As(V) determined by species-selective sampling as described in the sections 2.7.2 and 4.2.1.2. Thermodynamic data for gaseous As species were obtained from Spycher and Reed (1989) and Pokrovski et al. (2002) ( Appendix 26).

4.3 Results

4.3.1 Cluster analysis

Cluster analysis was conducted for both water and gas samples. It is discussed briefly in the follow-ing before the clusters obtained are used for the description of water and gas chemistry in the 5 study areas in section 4.3.2.

4.3.1.1 Clustering of the water samples

By cluster analysis 60 water samples from 5 study areas were divided in 14 subgroups. A logical subdivision in two major groups with 7 subgroups each was found with the cluster methods of Ward´s ( Appendix 11) and the Average Linkage within Groups method ( Appendix 12). Compared to those two methods, other clustering methods tested (section 4.2.3.4) were inferior and yielded many subgroups with only 1 or 2 single water samples without an overall context. Based on the dendrogram derived by Ward´s automatic cluster method and close examination of hydrogeo-chemical data, a few samples were manually re-assigned to different subgroups ( Appendix 11). In the Kruskall-Wallis test, the derived optimized cluster analysis showed significant overall differ-ences between the subgroups for all variables on a significance level of 0.1%, except for Co (0.6 %), Cd (0.7%), Pb (2.1%), and N(V) (4.2%). No significant overall difference was found for Cu ( Appendix 13). The results are discussed based on the cluster membership of the optimized cluster analysis of water samples.

The two major groups found during cluster analysis are sulfate-dominated waters with a low pH outside the caldera, mainly along the Norris-Mammoth corridor (33 samples) compared to chlo-ride-dominated waters with near-neutral pH closer to the caldera´s margin at Norris Geyser Basin and Gibbon Geyser Basin and, inside the caldera, at Lower Geyser Basin (27 samples, Figure 44).

Significant differences between the two groups also exist for temperature, Na, Cl, Br, F, B, and Li that show a direct positive correlation to pH as well as for S(VI), Mg, Al, Mn, and Fe that show an inverse correlation to pH ( Appendix 14). These two different groups correspond to the type I and type II waters discussed in section 4.1.1.5. The ternary diagram Cl-SO4-HCO3 in Figure 45 shows their origin from heated steam (type I) or deep thermal sources (type II).

Results 117

Figure 44 Distribution of type I steam-heated waters mainly along the Norris-Mammoth corridor to the north (33 samples, white trian-gles) and type II deep thermal waters closer to the caldera rim at Norris Geyser Basin (Ragged Hills) and Gibbon Geyser Basin as well as inside the caldera at Lower Geyser Basin (27 samples, black squares)

Figure 45 Ternary diagram Cl-SO4-HCO3 showing the origin of type I waters from heated steam and type II waters from deep thermal sources (after the classification scheme of Giggenbach and Goguel 1989)

Ternary diagram Cl-SO 4-HCO 3

Type I sulfate-dominated water (group 1-7)

Type II chloride-dominatedwater (group 8-14)

SO42-

Cl-

HCO 3 -

steam-heated waters

peripheralwaters

deep thermalwaters

12 5

10

4

5km 1 2 3 40 Lower Geyser Basin

3

Gibbon Geyser Basin 6

20

Nymph Lake Hazle Lake

Norris Geyser Basin

Caldera

N

Type I steam-heated waters

Type II deep thermal waters

Mammoth

West Yellowstone

Caldera

Numbers behind the symbols = number of samples


In a ternary diagram based on Na/K and K/Mg geothermometers (Giggenbach 1988) type I waters are classified as “immature waters”, while type II waters are in “partial” to “full equilibrium” with respect to the mineral system albite-potassium feldspar-muscovite-clinochlore-silica ( Figure 46). Saturation indices for these mineral phases were calculated by hydrogeochemical modeling to check the general validity of the simplified concept ( Appendix 15). Quartz is supersaturated in all samples, SiO2(a) in most, while Mg-clinochlore is undersaturated in all samples but one (YNPLG02-2W). Except for one sample (YNPHL02-1W), saturation and supersaturation of albite, potassium feldspar, and muscovite occurred as predicted only in samples characterized as being in “full equilibrium”. There are, however, also 6 sites classified as being in “full equilibrium” where albite, potassium feldspar, and muscovite are undersaturated. Thus, the simplified clustering was accepted as a first approximation to find water samples that show saturation or supersaturation with regard to the feldspar-mica system even though the classification of “full equilibrium” might be false in terms of thermodynamics for some of the samples. For the samples in full equilibrium, reservoir temperatures were calculated from the Si-, the Na-K-, the K-Mg-, and Na-K-Ca-geothermometers ( Appendix 16).

Figure 46 Ternary diagram Na-K-Mg showing type I sulfate-dominated, mostly immature waters, and type II chloride-dominated waters in partial or full equilibrium with the mineral system albite-potassium feldspar-muscovite-clinochlore-silica (equilibrium line calculated for 75 to 350°C in intervals of 25°C according to the equation for Na/K geothermometer by Arnórsson et al. 1998, cited in Arnórsson 2000 and for K/Mg geother-mometer by Giggenbach 1988)

The positive correlation between As and Cl discussed in general in section 4.1.1.5 was confirmed for the 60 water samples by bivariate correlation analysis on a significance level of 1% ( Appendix 14). Concentrations of As are generally higher in type II than in type I waters ( Figure 47). While As occurs almost exclusively as As(III) in type II waters, an average of 30% As(V), MMAA, and

Ternary diagram Na-K-Mg

Type I sulfate-dominated water (group 1-7)

Type II chloride-dominated water (group 8-14)

K/100

Na/1000

SQR(Mg)

fullequilibrium

partialequilibrium

immature waters

Results 119

DMAA can be found in type I waters. The significant difference between the two main types for the variables mentioned in this paragraph was confirmed by a Mann-Whitney test ( Appendix 17).

Figure 47 Positive correlation and significant trend of As and Cl concentrations due to similar behavior dur-ing steam-water separation; steam-heated type I waters show lower concentrations of As and Cl compared to deep thermal type II waters*

*The two samples of subgroup 9 (YNPRH05-1W, 3W with Cl: 399 and 385 mg/L, As: 6940 and 11085 µg/L) were ex-cluded from the graph and the calculation of the regression equation

The 7 subgroups each for type I and type II will be discussed in section 4.3.2.1 and section 4.3.2.2 respectively.

4.3.1.2 Clustering of the gas samples

For clustering the gas samples, the 55 NaOCl samples taken by gas hoods directly above hot springs (setup 1 and 2 in Figure 42) were considered. Significant concentrations of volatile metal-lics were found in the NaOCl trapping solution for As (from GF-AAS), Al, B, Ba, Cu, Fe, K, Li, Si, Sr, and Zn from ICP-AES, and S from IC determination as well as in the NaOH trapping solution for Cl. Background values of the NaOCl and NaOH trapping solution were considered (see digital database, Appendix 9).

The gas samples could not be subdivided in the same 14 subgroups as described for the water sam-ples in section 4.3.1.1. A Kruskall-Wallis test showed that the differences between the assumed water subgroups are not significant for none of the detected volatile elements (section 4.2.3.4). For 38 of the 55 gas samples water chemistry data was available for the same sampling time. The com-

Cl-As-correlation

0

500

1000

1500

2000

2500

3000

3500

0 100 200 300 400 500 600 700 800 900

Cl concentrations [mg/L]

As c

once

ntra

tions

[µg/

L]

Type I sulfate-dominated water (group 1-7) Type II chloride-dominated water (group 8-14)

y = 4.219⋅x + 91.768 r2 = 0.8196


bined data of gas and water analyses were used to conduct a hierarchical cluster analysis by Ward´s method ( Appendix 18). A logical subdivision in 6 subgroups was found and confirmed to yield significant differences between the subgroups for volatile As, Al, Fe, Li and all water parameters except for N(V), Ca, Cu, Sr, S(-II), and TIC on a significance level of 5% ( Appendix 19). The 17 gas samples for which no contemporaneous water samples existed were assigned to the 6 sub-groups based on knowledge of the sampling locations and the gas data. Considering the 55 gas samples only, and no water data, the 6 subgroups show significant differences for volatile As, Al, B, Ba, Fe, K, Li, and S on a significance level of 5% ( Appendix 20).

A significance level of 5% was considered acceptable for the cluster analysis of the gas samples compared to a significance level of 0.1% for the water samples (section 4.3.1.1). The significance level of 5% takes into account that volatile metallics are subject to changes in even more parame-ters than the water chemistry and generalizations are harder to make. The assignment of a sampling site to a certain subgroup based on its gas chemistry might be more sensitive to seasonal changes than based on its water chemistry. While 3 of the 6 gas subgroups contain exclusively gas samples over type I waters and 2 exclusively gas samples over type II waters, there is one subgroup with 3 samples both over type I and type II waters. This subgroup shows the highest volatile As concen-trations and supports the hypothesis that temporal changes can yield a completely different gas chemistry for one sampling site compared to the overall chemistry of its corresponding water phase.

In the following, the water and gas chemistry for the sites sampled will be discussed in two sections corresponding to the two main types of steam-heated (section 4.3.2.1) and deep thermal waters (section 4.3.2.2). Within the description of the individual study areas reference will be taken to the affiliation of sampling sites to individual subgroups found by water or gas cluster analysis.

4.3.2 Description of water and gas chemistry

4.3.2.1 Type I steam-heated waters

All of the type I waters are from outside the caldera, about half of them from sampling areas north of the Norris Geyser Basin along the Norris-Mammoth corridor. Seven subgroups can be distin-guished within the type I waters that are, except for a few samples, all steam-heated, immature waters ( Figure 48; Figure 49). The assignment of individual water samples to the 7 subgroups is given in Appendix 21. Figure 50 to Figure 53 show the main differences in the mean values of pH, conductivity, temperature, S(VI), Cl, dissolved total As and As species concentrations for the 7 water subgroups. Compared to type II waters As concentrations are low. They correlate with the low Cl concentrations ( Figure 50). High S(VI) concentrations lead to low pH ( Figure 51; Figure 53).

Results 121

Figure 48 Ternary diagram Cl-SO4-HCO3 showing the origin of the 7 subgroups of type I steam-heated wa-ters (modified after the classification scheme of Giggenbach and Goguel 1989, HCO3

- + 50 instead of HCO3-; for

subgroup affiliation see Appendix 21)

Figure 49 Ternary diagram Na-K-Mg showing the subgroups of type I sulfate-dominated waters that are mostly immature with respect to the mineral system albite-potassium feldspar-muscovite-clinochlore-silica (equi-librium line calculated for 75 to 350 °C in intervals of 25°C according to the equation for Na/K geothermometer by Arnórsson et al. 1998, cited in Arnórsson 2000 and for K/Mg geothermometer by Giggenbach 1988; for sub-group affiliation see Appendix 21)

Ternary diagram Cl-SO 4 -HCO3

Type I sulfate- dominated water

Low Cl hot springs

Frying Pan Spring September

2003 fumaroles Nymph Lake

Lakes

Hot springs Gibbon GB

"Verde Spring"

"Milky Way" - "Kaolin Spring"

SO42-

Cl -

HCO 3 - + 50

deep thermalwaters

steam-heatedwaters

Ternary diagram Na-K-MgType I sulfate-

dominated water

Low Cl hot springs



Lakes


"Verde Spring"


K/100

Na/1000

SQR(Mg)

fullequilibrium

partialequilibrium

immature waters


0

200

400

600

800

1000

1200

1400

1600

1800

2000

As

conc

entr

atio

n [µ

g/L

]

DMAAMMAAAs(V)As(III)

Cl-As correlation Type I sulfate-dominated water

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0 50 100 150 200 250 300


As c

once

ntra

tions

[µg/

L]

Low Cl hot springs



Lakes


"Verde Spring"


Figure 50 Positive correlation of As and Cl in steam-heated type I waters (subgroup affiliation: Appendix 21)

Figure 51 Differences in S(VI) and Cl concentrations for the 7 subgroups of steam-heated type I waters with a clear pre-dominance of S(VI)

Figure 52 Differences in dis-solved As species for the 7 sub-groups of type I waters; DMAA predominates only in subgroup 4 Lake samples

Subgroup No. 1 2 3 4 5 6 7

0

200

400

600

800

1000

1200

1400

1600

1800

Cl /

S(V

I) c

once

ntra

tions

[mg/

L]

S(VI)

Cl

Subgroup No. 1 2 3 4 5 6 7

Results 123

0

500

1000

1500

2000

2500

3000

3500

4000

4500

Low Cl hotsprings

Frying PanSpring

September

2003fumaroles

NymphLake

Lakes Hot springsGibbon GB

"VerdeSpring"

"MilkyWay" -"KaolinSpring"

cond

uctiv

ity [µ

S/cm

]

0

10

20

30

40

50

60

70

80

90

Tem

pera

ture

[°C

]

conductivity

Temperature

Figure 53 Differences in temperature, conductivity, and pH for the 7 subgroups of type I waters

Considering their gas chemistry, the samples of the 7 water subgroups can be subdivided into 4 gas subgroups. Figure 54 shows the differences of their mean volatile As, SiO2, and S concentrations. For the assignment of individual gas samples to the gas subgroups see Appendix 21.

Figure 54 Volatile As, SiO2, and S concentrations for the 3 gas subgroups A, B, C of type I waters; subgroup D also contains one sample of type II water that was not considered in the figure above: D(I); values are standard-ized to a range of 0 to 100 % (= highest concentration for each volatile species); mean value marked by symbol and number for mean concentration in mg/m3; bars represent maximum and minimum concentrations

pH = 2.6

pH = 2.4 pH = 2.0

pH = 3.6

pH = 3.4

pH = 2.3

pH = 2.7

Volatile As

28.3

8.0

103

14.70

10

20

30

40

50

60

70

80

90

100

A B C F

Volatile S iO2

3230

1000

1820

580

A B C F

Volatile S

46703800

6700

2850

A B C F

%

Subgroup No. 1 2 3 4 5 6 7

water 3 1, 2, 5, 7 4, 6 1, 7 [14] subgroup

gas / D(I) D(I) D(I)


Nymph Lake / Frying Pan Spring The area at the northern side of Nymph Lake with the new fumaroles appearing during the winter of 2002/2003 (section 4.1.2.1; Figure 35; Figure 55) shows the highest conductivities (3860 and 4190 µS/cm), the highest S(VI) concentrations (1.3 and 1.9 g/L) and the lowest pH (2.0 and 1.9) of all samples ( Figure 51; Figure 53). Judging by the cluster analyses of both gas and water data ( Appendix 11; Appendix 12; Appendix 18) these two samples (YNPNL05 and 06) have the least similarity with any other site sampled. Concentrations of dissolved Al (48 and 70 mg/L), B (2.7 and 3.8 mg/L), Fe (30 and 34 mg/L), SiO2 (304 and 374 mg/L), Pb (8 µg/L), Zn (0.051 and 0.058 mg/L) as well as total Hg (600 ng/L) and methylated Hg (0.55 ng/L) are significantly increased compared to the other samples within the subgroup. For the June sample, rare earth elements were analyzed. Almost all rare earth elements showed higher concentrations than for the other sites sam-pled, especially U (1.8 µg/L), Th (15 µg/L), Ce (40 µg/L), and La (20 µg/L). From June to July, pH decreased whereas temperature, conductivity and S(VI) concentrations increased. In contrast to the high concentrations of dissolved species, almost all volatile species at YNPNL05 and 06 (gas sub-group A; Figure 54) show concentrations lower than at all other sampling sites. Volatile As concentrations range from 7.0 to 8.9 mg/m3 (quantification is discussed in detail in section 4.3.6). Volatile B concentrations are below detection limit, Li (2.8 mg/m3), Sr (2.3 mg/m3), SiO2 (580 mg/m3) and S concentrations (2.8 mg/m3) are among the lowest concentrations sampled at all sampling sites.

Figure 55 Sampling sites at Nymph Lake: 1 = Lake sample; 2, 3, 4 (Frying Pan spring) = low Cl hot springs and fumaroles (September samples from 3 and 4 belong to another subgroup); 5, 6 = area of the new fumaroles; line is clearly visible by the dead trees; site symbols correspond to subgroup symbols used in the previous figures

The 3 sites sampled at Frying Pan Spring (YNPNL02, 03, 04; Figure 35 and Figure 55) belong to the subgroup 1 of hot springs and fumaroles with the lowest Cl concentrations (on average 5 mg/L; Figure 51). This subgroup also shows the lowest As concentrations with an average of 80 µg/L. Methylated Hg concentrations determined in June at NL02 (9.04 ng/L), and NL04 (6.3 ng/L) are

6 5

1

2 3 4

N

Results 125

the highest found throughout the sampling campaigns. Total Hg concentrations of 72 ng/L at NL04 are lower compared to other sampling sites. Volatile As concentrations with an average of 11.5 mg/m3 are the second lowest of all areas sampled after the Nymph Lake fumaroles described before (subgroup B; Figure 54).

During the third sampling campaign in September, YNPNL03 and 04 showed significantly in-creased conductivities, caused mainly by increased concentrations of H+ and S(VI). Some trace elements like Al, Mn, Fe (at NL04 only), Sr, and Zn were increased, too. Arsenic concentrations decreased with decreasing pH. A trend could already be observed during the first two sampling campaigns with conductivities increasing from June to July ( Figure 56). The relative distance in cluster analysis between the samples YNPNL03-3W and 04-3W and the samples from subgroup 1 required their separation in an own subgroup 2 ( Appendix 11; Appendix 12). Volatile As concen-trations increased from 4.3 mg/m3 in June and 3.8 mg/m3 in July to 90 mg/m3 in September (NL04). Such differences in gas and water chemistry between sampling from early summer to late autumn, detected also at other sampling sites (see below), are assumed to be no artifacts but an expression of the seasonal “disturbances” discussed in section 4.1.2.3 for Norris Geyser Basin. Interestingly, YNPNL02 (all sampling campaigns in subgroup 1), shows an inverse trend with de-creasing conductivities, decreasing S(VI) concentrations, and an increasing pH from June through July to September ( Figure 56). The number of samples is unfortunately too low to confirm a sea-sonal shift in activity between the northeastern and the southwestern part of the Frying Pan Spring area.

1596

1368

443

965

176

440

2.452.562.5

0.330.280.19

0

500

1000

1500

2000

2500

3000

3500

YNPNL02-1W YNPNL02-2W YNPNL02-3W

cond

uctiv

ity [µ

S/cm

] and

S(V

I) [m

g/L

0

0.5

1

1.5

2

2.5

3

3.5

pH /

As [

mg/

L]

conductivityS(VI)pHAs


Figure 56 Seasonal trends at Frying Pan Spring from June through July to September with increasing conductivity and S(VI) concentrations, and decreasing pH and As concentrations at YNPNL03 and YNPNL04 and vice versa at YNPNL02

800

461

1500

124

517

257

3.01

2.37

2.68

0.270.26

0.50

0

500

1000

1500

2000

2500

3000

3500


cond

uctiv

ity [µ

S/cm

] and

S(V

I) [m

g/L

0

0.5

1

1.5

2

2.5

3

3.5

pH /

As [

mg/

L]


375

1663 1725

115

539406

3.04

2.332.33

0.180.210.37

0

500

1000

1500

2000

2500

3000

3500


cond

uctiv

ity [µ

S/cm

] and

S(V

I) [m

g/L

0

0.5

1

1.5

2

2.5

3

3.5

pH /

As [

mg/

L]


Results 127

The only sample taken at Nymph Lake (YNPNL01) belongs to subgroup 4 of the lake samples and will be discussed with the samples from Hazle Lake.

Hazle Lake Two samples from the Hazle Lake shores (YNPHL01 and 02) constitute the subgroup 4 “lake sam-ples” together with YNPNL01. Characteristic are the low temperatures (average 24°C; Figure 53), the highest pH in their group of sulfate-dominated waters (on average 3.6, up to 6.0 at HL02) and a higher than average concentration of Ca (13 mg/L) and TIC (13 mg/L). At YNPHL01 1.6 ng/L methylated Hg were detected. Total As concentrations were higher than in the Nymph Lake / Fry-ing Pan Spring area with an average of 280 µg/L and - in contrast to all other sampling sites - significant amounts of DMAA ( Figure 57). Volatile As concentrations (subgroup C; Figure 54) are also higher. However, no significant correlation with volatile As was found neither for dissolved total As nor DMAA. Bivariate correlation analysis revealed that only dissolved As(III) correlates with volatile As ( Appendix 23). Lower volatile Si concentrations for the lake samples compared to the hot springs from Frying Pan Spring correlate with lower temperatures ( Figure 53; Figure 54; Appendix 23). For the lower volatile S concentrations no explanation could be found, because vola-tile S does not correlate with any other gaseous or aqueous parameter ( Appendix 22; Appendix 23).

Comparing the three sampling campaigns from June, July, and September at HL01, increasing con-ductivities (1030 - 1456 - 1890 µS/cm) reflect increasing concentrations of Cl, S(VI), Ca, K, Mg, and Na as well as F, Br, B, Li, and Sr. Total As concentrations increase from 197 to 414 and 536 µg/L. Dimethylated As predominates in June and September, As (III) in July ( Figure 57). The Fe(II) increases from 88 to 94 and 96% and TOC from 1.8 to 3.3 and 6.4 mg/L.

The small hot spring at the northern side of the lake about 3 m above lake level (YNPHL03) be-longs to the subgroup of low Cl hot springs and shows a clear predominance of As(III) like the hot springs at Frying Pan Spring ( Figure 57).


Figure 57 Aerial overview of Hazle Lake with the lake samples YNPHL01 and YNPHL02 and the low Cl hot spring YNPHL03; pie charts show that significant amounts of dimethylated As are found in the lake samples only, while As(III) clearly predominates in the hot spring sample YNPHL03 (concentrations in µg/L); numbers in pho-tos correspond to bold numbers in sampling code

226

223

155

1

2 3

3

1

N

49

39

36

62

21

2

124391

1 14

209

45

13

85

YNPHL01-1W

YNPHL03YNPHL02

YNPHL01-2W

YNPHL01-3W

As(III) As(V)

MMAA DMAA

Results 129

Norris Geyser Basin (Ragged Hills) The only sites in the Ragged Hills area that belong to type I steam-heated waters are “Verde Pool” ( Figure 38) and the “Milky Way” - “Kaolin Spring” complex that is connected by a discharge breakthrough from “Kaolin Spring” since the winter 2002/ 2003 ( Figure 39).

The highest As concentrations (1500 µg/L) in steam-heated waters with a predominance of As(III) (76%) are found in the “Verde Pool” subgroup ( Figure 50; Figure 52). Conductivities of about 2500 µS/cm are second only to samples from the 2003 fumaroles at Nymph Lake ( Figure 53). Av-erage concentrations of Cl are 200 mg/L, S(VI) 500 mg/L, and SiO2 306 mg/L. The pH is low with an average of 2.3. Temperatures of 45°C are the second lowest apart from the lake samples. Ac-cordingly, the gas chemistry is similar to that of the lake samples described before with high vola-tile As and low volatile Si and S concentrations ( Figure 54). Methylated Hg concentrations of 0.03 and 0.04 ng/L for YNPRH03-1W and 04-1W respectively are the lowest sampled during this field trip. From June through July to September pH decreased while conductivity and S(VI) concentra-tions increased at YNPRH04 as observed for YNPHL01, NL03, 04, and 05. Total As concentra-tions peaked in July with 1968 µg/L compared to 1895 µg/L in June and 1778 µg/L in September. The As(III) fraction increased from 70 to 77 and 83% throughout September.

With concentrations of 200 mg/L for both Cl and S(VI) ( Figure 51), samples from the “Milky Way” - “Kaolin Spring” complex (subgroup 7) plot between the fields for deep thermal waters and steam-heated waters in Figure 48. Average As concentrations are 660 µg/L with a clear predomi-nance of As(III) (81%). With 490 ng/L (YNPRH18-4W) and 600 ng/L (YNPRH01-1W) total Hg concentrations are among the highest found in the Park. Methylated Hg concentrations are lower, being 0.22 and 0.47 ng/L respectively. The sampling site YNPRH01 showed a slight decrease in conductivity from June through July to September (1608 - 1512 - 1440 µS/cm) mainly from de-creasing Cl concentrations (245 to 209 mg/L). Sulfate concentrations stayed constant at 205 mg/L.

Upon close observation a linear structure seems to connect those two zones in the aerial photograph of the Ragged Hills area in Figure 58. The most prominent features along that “lineament” are a red alteration zone at the southern side of ”Verde Pool“, linear clearings through the sparsely wooded hill side east of “Verde Pool“ and a small hot spring west of the “Milky Way” - “Kaolin Spring” complex. Unfortunately no information exist about the mineral composition or genesis of the red alteration zone nor about the hydrogeochemistry of the small hot spring that could support the hy-pothesis that the similar hydrogeochemistry between “Verde Pool” and the “Milky Way” - “Kaolin Spring” complex might be explicable by their location on one fracture zone providing ascent paths for geothermal gases.


Figure 58 Aerial photograph of the Ragged Hills area with “Verde Pool” (YNPRH03, 04) and the “Milky Way” - “Kaolin Spring” complex (YNPRH01, 18, 22); a linear structure seems to connect both areas (dashed white line arbitrarily drawn a few millimeter below the structure shows its heading; AZ = alteration zone, LC = linear clearings; HS = hot spring); numbers in photo correspond to bold numbers in sampling code

Gibbon Geyser Basin (Geyser Springs group) The samples from subgroup 5 (Gibbon Geyser Basin GG04 and GG08; Figure 40) are closest to equilibrium with respect to the mineral system albite-potassium feldspar-muscovite ( Figure 49). They have the highest pH apart from the lake samples with an average pH of 3.4, Cl concentrations of 150 mg/L and S(VI) concentrations of 320 mg/L ( Figure 51). Distinctive are their high tempera-tures of 83°C ( Figure 53), the high Na concentrations of 190 mg/L, and the high F concentrations of 4.5 mg/L (average each). Arsenic concentrations are on average 630 µg/L with a clear predomi-nance of As(III) (82% on average). No significant differences were observed for the three sampling campaigns at GG04, except for As that shows higher concentrations during the second sampling campaign (823 µg/L) compared to the first and forth (580 and 528 µg/L) and an increasing As(III) fraction of 74 to 79 to 84% from June through July to October.

4.3.2.2 Type II deep thermal waters

Type II waters only include samples from Norris Geyser Basin, Gibbon Geyser Basin and Lower Geyser Basin, no samples from the Nymph Lake and Hazle Lake area along the Norris-Mammoth corridor. Seven subgroups can be distinguished within the deep thermal type II waters that are mostly in partial to full equilibrium with regard to the mineral system albite-potassium feldspar-muscovite ( Figure 59; Figure 60). The assignment of individual water samples to the 7 subgroups is given in Appendix 21.

3 4

1 18 22

N

LCHS

LCAZ

?

?

?

Results 131

Ternary diagram Cl-SO4-HCO3

Type II chloride-dominated water

Lower GB alkaline springs

Ragged Hills No.5

Ragged Hills No.11

Near neutral springs Gibbon GB

Warm springs Ragged Hills

Shallow hot springs RH

Deeper hot springs RH

SO42-

Cl-

HCO3- + 50

steam-heatedwaters

deep thermalwaters

Figure 59 Ternary diagram Cl-SO4-HCO3 showing the origin of the subgroups 8-14 of type II deep thermal waters (modified after the classification scheme of Giggenbach and Goguel 1989, HCO3

- + 50 instead of HCO3-; for

subgroup affiliation see Appendix 21)

Figure 60 Ternary diagram Na-K-Mg showing the subgroups of type II chloride dominated waters mostly in partial or full equilibrium with the mineral system albite-potassium feldspar-muscovite-clinochlore-silica (equilib-rium line calculated for 75 to 350 °C in intervals of 25°C according to the equation for Na/K geothermometer by Arnórsson et al. (1998, cited in Arnórsson 2000) and for K/Mg geothermometer by Giggenbach (1988); for sub-group affiliation see Appendix 21)

Ternary diagram Na-K-MgType II chloride- dominated water


Ragged Hills No.5

Ragged Hills No.11





K/100

Na/1000

SQR(Mg)

fullequilibrium

partialequilibrium

immature waters


Figure 61 to Figure 64 show the main differences in mean values of pH, conductivity, temperature, S(VI), Cl, dissolved total As and As species concentrations for the 7 water subgroups.

Cl-As correlation Type II chloride-dominated water

0

500

1000

1500

2000

2500

3000

3500

0 100 200 300 400 500 600 700 800


As c

once

ntra

tions

[µg/

L]


Ragged Hills No. 11





Figure 61 Positive correlation of As and Cl concentrations in deep thermal type II waters (the two samples of subgroup 9 (YNPRH05-1W, 3W with Cl: 399 and 385 mg/L, As: 6940 and 11085 µg/L) are excluded from the graph (outlier); for subgroup affiliation see Appendix 21)

0

500

1000

1500

2000

2500

3000

3500

4000

Lower GBalkalinesprings

RaggedHills No. 5

RaggedHills No. 11

Near-neutral hot

springsGibbon GB

WarmspringsRaggedHills

Shallow hotspringsRaggedHills

Deeper hotspringsRaggedHills

cond

uctiv

ity [µ

S/cm

]

0

10

20

30

40

50

60

70

80

90

100T

empe

ratu

re [°

C]

conductivityTemperature

Figure 62 Differences in temperature, conductivity, and pH for the 7 subgroups of type II waters (for sub-group affiliation see Appendix 21)

pH = 7.7

pH = 2.9

pH = 2.3 pH = 6.4

pH = 2.9

pH = 3.2 pH = 3.8

Subgroup No. 8 9 10 11 12 13 14

Results 133

Figure 63 Differences in S(VI) and Cl concentrations for the 7 subgroups of deep thermal type II waters with a clear pre-dominance of Cl

Figure 64 Differences in dissolved As species for the 7 subgroups of type II waters with an overall clear predominance of As(III)

Volatile As

0

10

20

30

40

50

60

70

80

90

100

Figure 65 Volatile As, SiO2, and S concentrations for the 2 gas subgroups E and F of type II waters; subgroup D contains only one sample of type II waters: D(II); type I waters are not considered ( Figure 54); values are stan-dardized to a range of 0 to 100 % (= highest concentration for each volatile species); mean value marked by sym-bol and number for mean concentration in mg/m3; bars represent maximum and minimum concentrations; <D.L. = below detection limit

Volatile SVolatile S iO2

0

200

400

600

800

1000

1200C

l / S

(VI)

con

cent

ratio

ns [m

g/L]

S(VI)

Cl

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

As c

once

ntra

tion

[µg/

L]

DMAA

MMAA

As(V)As(III)

Subgroup No. 8 9 10 11 12 13 14

Subgroup No. 8 9 10 11 12 13 14

water 8, 11, 14 9,10 [1, 7,] 14 subgroup

gas /

%

E F D(II) E F D(II) E F D(II)

10.8

89

130

2170

4910 3130 1090

2050

< D.L.


Considering their gas chemistry, the samples of the 7 water subgroups can be subdivided into the gas subgroups E (Lower Geyser Basin alkaline hot springs; near neutral pH hot springs at Gibbon Geyser Basin and deeper hot springs Ragged Hills) and F (Ragged Hills No. 5 and No. 11) ( Appendix 21). Figure 65 shows the differences of the mean values of volatile As, SiO2, and S con-centrations for these 3 subgroups. Subgroup D with the highest volatile As concentrations contains only one type II water sample (YNPRH20), all others are type I as shown in Figure 54. Norris Geyser Basin (Ragged Hills) Five different subgroups of type II waters could be distinguished at Ragged Hills ( Figure 66).

Figure 66 Ragged Hills (aerial photograph from Becker 2004) with sampling locations of the type II waters; numbers in photo correspond to bold numbers in sampling code

The majority of type II waters at Ragged Hills belongs to two subgroups (subgroup 13 with 6 cases and subgroup 14 with 8 cases) with the highest Cl concentrations sampled ( Figure 61). Subgroup 13 contains mainly smaller features of only a few centimeters water depths like “Milky Way II” or “Dragon Spring”, subgroup 14 mainly the larger structures like the “Persnickety” - “Titanic” -

19

20 21 25

24

28

15

16

13

14

10

17 8

5

11

Shallow hot springs Ragged Hills YNPRH17-4W; YNPRH24-4W; YNPRH25-4W; [YNPRH26-4W]; [YNPRH27-4W]; YNPRH28-4W Deeper hot springs Ragged Hills YNPRH08-1W,4W; YNPRH10-4W; YNPRH13-3W; YNPRH14-3W; YNPRH19-4W; YNPRH20-4W

Ragged Hills No. 5 YNPRH05-1W, 3W

Ragged Hills No. 11 YNPRH11-3W

Warm springs Ragged Hills YNPRH15-3W; YNPRH16-4W

26

27

N 0 20 40 60 80 100 m

Results 135

“Life Boat” complex, “Elk Geyser”, “Rainbow Growler” and Crystal Springs. The bottom of these larger structures is often not visible, thus exact water depths can not be defined. Comparing the smaller, shallower structures with the larger, deeper structures, average conductivities (2038 and 2168 µS/cm), pH (3.2 and 3.8), average concentrations of Si (268 and 382 mg/L), Cl (505 and 603 mg/L), Na (278 and 335 mg/L), Ba (0.16 and 0.05 mg/L), and Sb (75 and 130 µg/L) are generally lower ( Figure 62; Figure 63; Figure 67). Average S(VI) concentrations are higher in the shallower structures compared to the larger ones (94 and 69 mg/L; Figure 63). From the two samples taken for Hg YNPRH08 had 0.35 ng/L methylated Hg and 170 ng/L total Hg, YNPRH20-4W only 0.04 ng/L methylated Hg and 28 ng/L total Hg. The gas chemistry of the deeper hot springs is similar to that of hot springs at Gibbon Geyser Basin and Lower Geyser Basin (see below).

Figure 67 Larger deeper geothermal features like the “Life Boat” – “Titanic” – “Persnickety complex” and “Elk Geyser” show higher electrical conductivities (EC), SiO2, Cl, and Na concentrations (units on primary y-axis in µS/cm and mg/L) and lower Ba concentrations (units on secondary y-axis in mg/L) compared to the smaller and shallower structures like “Dragon Spring” and “Milky Way II” The two samples from the subgroup of “warm springs” in the western part of the Ragged Hills area are characterized by lower than average temperatures (37.8 and 38.9 °C for YNPRH15-3W and 16-4W respectively; Figure 62). Remarkable are their high average Ba concentrations of 0.24 mg/L. In mid September, the hot spring YNPRH15 showed a smeary bio-film on its surface with single flakes of foam floating on top ( Figure 68). After 2 weeks this bio-film disappeared and the hot spring´s water was clear again.

“Dragon Spring“ “Milky Way II“

“Elk Geyser“ ”Life Boat“ - ”Titanic“

0

50 0

10 0 0

150 0

2 0 0 0

2 50 0

0 .0 0

0 .0 5

0 .10

0 .15

0 .2 0

0 .2 5

0

50 0

10 0 0

150 0

2 0 0 0

2 50 0

0 .0 0

0 .0 5

0 .10

0 .15

0 .2 0

0 .2 5

0

50 0

10 0 0

150 0

2 0 0 0

2 50 0

0 .0 0

0 .0 5

0 .10

0 .15

0 .2 0

0 .2 5

0

50 0

10 0 0

150 0

2 0 0 0

2 50 0

0 .0 0

0 .0 5

0 .10

0 .15

0 .2 0

0 .2 5

EC

SiO2

Ba

Cl Na

EC

SiO2

Ba

Cl Na

EC

SiO2 Ba Cl

Na

EC

SiO2 Ba Cl

Na


Figure 68 Warm spring YNPRH15 mid of September with white foam flakes on its surface

The highest As concentrations (6.9 and 11 mg/L; Figure 64) were found at a small hot spring of approximately 1 m diameter (YNPRH05; Figure 69). Arsenite predominates at this site with 89 and 94% for the June and September sampling respectively. In Figure 60 the sampling site plots in the field between immature waters and waters in partial equilibrium with respect to the mineral system albite-potassium feldspar-muscovite. It shows low pH (2.9) and relatively high S(VI) concentra-tions (190 mg/L) within its group of chloride-dominated waters ( Figure 62; Figure 63).

Figure 69 Sampling YNPRH05, with 6.9 and 11 mg/L the hot spring with the highest dissolved As concentrations

YNPRH15

gas sampling site

N

YNPRH05

N

Results 137

The lowest pH (2.3) and the highest conductivity (3350 µS/cm) within the group of type II waters, shows YNPRH11, an ellipse shaped (1 to 2 m) hot spring that changed to mud pot activity during October 2003 ( Figure 62). With 541 mg/L it contains a significant amount of S(VI) besides 463 mg/L Cl ( Figure 63) and plots between deep thermal and steam-heated waters ( Figure 59). It is also the most immature water within its group ( Figure 60). Arsenic concentrations of 3.4 mg/L are the second highest after YNPRH05 ( Figure 61; Figure 64).

The two sampling sites with the highest total As concentrations, YNPRH05 and 11, belong to the gas subgroup F, characterized by the second highest volatile As concentrations with a mean of 89 mg/m3 ( Figure 65). The fact that these samples also show the highest dissolved As(III) fraction of all samples ( Figure 64) corresponds well with the observed correlation of dissolved As(III) and volatile As ( Appendix 23). The sites YNPRH05 and 11 are also the only ones, besides those of gas subgroup D, with significant concentrations of volatile Li (42 mg/m3) and B (138 mg/m3). Gibbon Geyser Basin (Geyser Springs group) Subgroup 11 contains the sites GG02, 06, and 07 at Gibbon Geyser Basin and RH09 at Ragged Hills. Their pH is near neutral (6-7). While all the samples from Gibbon Geyser Basin within this subgroup are in full equilibrium, YNPRH09 is only in partial equilibrium with respect to the min-eral system albite-potassium feldspar-muscovite ( Figure 60). Low concentrations of methyl Hg were detected in YNPRH09 (0.04 ng/L).

Like the deeper hot springs at Ragged Hills and Lower Geyser Basin, chloride-dominated hot springs at Gibbon Geyser Basin show the lowest volatile As concentrations of all samples with an average of 10.8 mg/m3 (gas subgroup E; Figure 65). A negative correlation with the highest dis-solved TIC concentrations (average 17 mg/L) was found to be significant on a level of 3.2% ( Appendix 23). Remarkable are also the low concentrations of volatile Al (200 mg/m3), Fe (14.5 mg/m3), and Ba (43 mg/m3) compared to other subgroups. They correlate with low volatile As con-centrations and dissolved Fe(III) concentrations close to detection limit (0.03 mg/L) or below (for correlation analysis replaced by 0.3 ⋅ detection limit). Lower Geyser Basin (Pocket Basin) All 3 samples taken from within the Yellowstone caldera, at Pocket Basin in the Lower Geyser Basin area ( Figure 41), form their own subgroup. They are the only samples with alkaline pH from 7.3 to 8.7 ( Figure 62). At 91°C they have the highest temperatures of all samples ( Figure 62), the highest F concentrations (25 mg/L), and the lowest S(VI) concentrations (34 mg/L) ( Figure 63). Within their group of type II waters, they have the lowest Cl concentrations (318 mg/L) and the lowest conductivities (1400 µS/cm). The water at all three sampling sites is in full equilibrium with respect to the mineral system albite-potassium feldspar-muscovite ( Figure 60). Total inorganic carbon concentrations of 32 mg/L indicate an influence of peripheral waters on the deep thermal waters ( Figure 59). Remarkable are the W concentrations (390 to 430 µg/L) being 1 to 2 orders of magnitude higher than in all other samples. Cluster analysis confirms the differences of this sub-


group from all other type II subgroups, containing only samples from Gibbon Geyser Basin and Ragged Hills. The gas chemistry is similar to deeper hot springs at Ragged Hills and chloride-dominated hot springs at Gibbon Geyser Basin as described above.

4.3.3 Modeling As species and mineral phases

Speciation calculations show that in accordance with other reports on geothermal As (Cleverley et al. 2003; Spycher and Reed 1989; Webster and Nordstrom 2003) H3AsO3(aq) is the predominant As species in almost all samples ( Appendix 24). A predominance of As(V) species is only found for samples from Hazle and Nymph Lake, identified in the speciation modeling as H2AsO4

-. Spe-cies separation showed that those pentavalent species are probably rather methylated than inorganic species ( Figure 57; Figure 70 left) but the organic species could not be considered in the speciation modeling because of a lack of thermodynamic data. Modeling confirms that the non-charged pen-tavalent complex H3AsO4

0 would make up only 20% on average of all As(V) species. Its concentra-tion is more or less equal to that of H2AsO4

- in 7 of 60 cases and it becomes the predominant As(V) species with 70 and 75 % only in the two samples from the Nymph Lake fumaroles (YNPNL05). The total share of inorganic As(V) in these two samples is, however, <1% ( Figure 70 right). This minimal occurrence of the non-charged As(V) complex is important to justify the application of the ion exchanger method with the problem of non-charged As(V) being determined erroneously as As(III) as discussed in section 4.2.3.1.

YNPHL01-3W

DMAA

MMAA

AsS(OH)(HS)-

H3AsO3

Figure 70 Left: Predominance of methylated species* in Hazle Lake, the share of non-charged H3AsO4 is significantly lower than in other geothermal samples; Right: Only in samples from the new fumarole area at Nymph Lake did the non-charged H3AsO4 predominate over H2AsO4

- but both As(V) species make up less than 1% of all As species, the non-charged As(III) species H3AsO3 predominates like in most geothermal waters * Share of methylated species derived from on-site separation as explained in the sections 2.7.2 and 4.2.1.2; individual inorganic As(III), As(III)-S, and As(V) complexes modeled with PhreeqC, for modeling methylated species were added to the As(V) species as explained in section 4.2.3.5

YNPNL05-1W

DMAA

MMAA

H3AsO4

H2AsO4-

H4AsO3+

H3AsO3

Results 139

In all samples with near neutral to alkaline pH and S(-II) concentrations above detection limit (YNPLG01-2W, 03-2W, YNPGG02-2W, 06-2W, 07-2W, 4W, YNPRH08-4W) AsS(OH)(HS)- was calculated in concentrations of several tens to hundreds µg As/L ( Figure 71 top left). At YNPLG02 with a pH of 8.7, the S(-II) concentrations are so high (3.75 mg/L) that polysulfides (S5

2-, S42-, and

S62-) form according to the speciation modeling, and the concentration of AsS(OH)(HS)- (20 µg/L)

is negligible ( Figure 71 top right). At Hazle Lake, low pH values of 3.0 to 3.1 inhibit the formation of AsS(OH)(HS)- at As concentrations of 200-540 µg/L, even though S(-II) concentrations are high (0.9-1.7 mg/L). The concentrations of AsS(OH)(HS)- are 1 to 10 µg/L only, H2S is the predominant S(-II) species. For three sites at Ragged Hills (YNPRH05-1W, 3W, YNPRH11 and 26) modeling showed AsS(OH)(HS)- (30 to 90 µg/L) at pH values as low as at Hazle Lake (2.3 to 3.0). Still, H2S is the predominant S(-II) species, but higher As concentrations (1.7 to 11 mg/L) lead to higher AsS(OH)(HS)- concentrations. The predominant As-S species is As3S4(HS)2

- (130 to 240 µg/L) that only occurs in these four samples ( Figure 71 left).

YNPLG01-2WMMAADMAA

H3AsO3

H2AsO3-

H2AsO4-HAsO4

2-

AsS(OH)(HS)-

Figure 71 Occurrence of AsS(OH)(HS)- at S2--rich hot springs of near neutral pH (YNPLG01, LG02 above); LG02: because of very high concentrations of S2- polysulfides form predominantly and AsS(OH)(HS)- concentrations are significantly de-creased; As3S4(HS)2

- occurs only in acid hot springs with high S2- and As concentrations of several mg/L (YNPRH26 left)

Supersaturation with regard to the As minerals listed in the WATEQ4F database was found only for 5 samples at 3 sites ( Appendix 25). At Hazle Lake the saturation index for orpiment is +0.4,

YNPLG02-2W

DMAA

MMAA

AsS(OH)(HS)-

HAsO42-

H2AsO3-

H3AsO3

YNPRH26-4WDMAAMMAA

H2AsO4-

AsS(OH)(HS)-

As3S4(HS)2-

H3AsO4

H3AsO3


+0.6 and +1.6 for the June, July, and September sampling respectively. In September, additionally amorphous As2S3 was slightly supersaturated (SI = 0.2). According to Cleverley et al. (2003) amorphous As2S3 can be observed in meta-stable equilibrium with orpiment in natural settings. A possible precipitation of orpiment at Hazle Lake is not as striking bright yellow as at “Orpiment Puddle” and “Orpiment Puddle Son” but the sides of the small geothermal feature are covered with a yellowish precipitation the color of which is mainly obscured by the color of the algae rich lake water ( Figure 72). No microprobe analysis have been conducted yet to confirm whether the precipi-tate is orpiment. The temperatures and S(-II) concentrations observed at Hazle Lake are lower (20-26°C and 0.9-1.7 mg/L) compared to “Orpiment Puddle” (43-53°C and 3.5-4.25 mg/L; Nordstrom et al. 2003), pH, conductivity, redox potential, and As concentrations are comparable. Like “Orpi-ment Puddle” the samples from Hazle Lake show an anomalously low As/Cl ratio compared to all other samples indicating As removal by precipitation ( Figure 50).

Figure 72 Left: Geothermal feature at the south side of Hazle Lake (YNPHL01); yellowish precipitations at the sides could indicate the occurrence of orpiment that is supersatured according to PHREEQC modeling ( Appendix 25); right: for comparison: the bright yellow “Orpiment Puddle Son” USGS sampling September, 4th, 2003 from left to right: Beate Böhme, Jim Ball, JoAnn Holloway, and Kirk Nordstrom

The other two samples that show supersaturation for an As mineral phase are YNPRH15 and 16, the warm springs at Ragged Hills ( Figure 62). Saturation indices for realgar are +2.4 and +1.8 re-spectively. The precipitation of realgar directly from solution at temperatures less than 100°C is not very commonly found (Nordstrom and Archer 2003). Also the samples do not show a depletion of As compared to Cl that could indicate mineral precipitation ( Figure 61). However, the precipitation of an As rich-phase was actually observed when taking a gas sample by passive sampling from the 4th of October until the 7th of October 2003 close to YNPRH15 ( Figure 68). During the 3 days of sampling a metallic precipitation formed under the gas hood ( Figure 73). Upon removal of the gas hood the precipitation vanished within half an hour. The experiment was repeated from 24th to 27th of October and the results were confirmed. Microprobe analysis indicated a material rich in Fe, S, and As that is amorphous to electron microscope diffraction (HRTEM and SEM).

Results 141

Figure 73 Precipita-tion of amorphous Fe, S, and As rich material at YNPRH15 underneath a gas hood (33 x 24 x 8 cm) installed for 3 days to sample volatile metal-lics on SPME

Unfortunately this site was not sampled for total volatile metallics but only species-selective with a SPME fiber. No As species were identified on the SPME fiber. This result could indicate a scav-enging effect by the mineral precipitation, but the reliability of reproducible As sorption on the SPME fibers is still questionable as discussed in the next section ( 4.3.4).

4.3.4 Volatile As speciation

No volatile As species were detected neither on the 20 Tenax nor on the 20 Carboxen 564 solid sorbents. This result might be attributed to the lower sensitivity of solid sorbents compared to SPME fibers, but also to longer storage times. Analyses were not performed until 9 to 15 days later since the solid sorbents had to be sent back to Germany for a GC-MS with a thermal desorption port.

From the 57 gas samples taken by SPME fibers (12 PDMS 100, 31 PDMS-CAR, 14 PDMS-CAR-DVB) 34 showed no volatile As. Four different volatile As species were identified in the remaining 23 samples. Chromatograms and mass spectra are presented in the digital database ( Appendix 9). Peak areas are indicated for internal comparison only, a quantification was not attempted as ex-plained in section 3.2.3.2. Three fibers not exposed to the geothermal gases but stored in the same way as the gas samples were used as blank values and did not show any volatile As peaks. One PDMS and 2 PDMS-CAR fibers broke in the field and no samples could be obtained. The volatile As species detected on SPME fibers did not show apparent differences between the 6 gas sub-groups.

The most predominant volatile As species identified in 22 samples is (CH3)2AsCl. It has only been reported from laboratory experiments (section 2.5.3.6). Far less common is (CH3)AsCl2 detected in 3 samples with a peak area about 20-40% smaller than that of (CH3)2AsCl ( Figure 74). Its detection in a natural aquatic environment is also a novelty. Retention times for (CH3)2AsCl and (CH3)AsCl2


are about 3.2 and 5.6 minutes respectively. The fully chlorinated AsCl3 with a boiling point of 130°C did not occur in any sample. The fully methylated (CH3)3As with a boiling point of 52°C (section 2.5.3.4) and short retention times of about 1.9 minutes was detected in 8 chromatograms. The fourth species positively identified in 7 samples is (CH3)2AsS(CH3) ( Figure 75; Kösters et al. 2003). It has the highest boiling point of the four As species. Its retention time is about 6.1 minutes and it always occurred together with large peak areas of (CH3)2AsCl and with one exception also of (CH3)3As. Its discovery is especially interesting because it has only been synthesized in the labora-tory and nothing is known about its abundance in the environment (section 2.5.3.5). No correlation between the occurrence of volatile chloroarsines and dissolved or gaseous Cl and between the oc-currence of volatile sulfurarsine and dissolved or gaseous S was found. Trimethylarsine was pref-erably detected at low total volatile As and S concentrations and low As(III) ratios in water. Be-cause of the lack of calibration standards neither peak heights nor peak areas were calibrated. Therefore a meaningful correlation analysis could not be performed.

In respect to the choice of SPME fiber material no significant sorption preference was found in contrast to the different sorption behavior observed in the laboratory (section 3.2.3.2; Figure 26). The only generalization that can be made is that volatile (CH3)2AsS(CH3) was exclusively detected on PDMS-CAR and PDMS-CAR-DVB, not on PDMS 100. The preferable sorption of (CH3)2AsCl and (CH3)AsCl2 on PDMS 100 and of (CH3)3As and (CH3)2AsS(CH3) on PDMS-CAR was only observed for the sampling site YNPLG02. At all other sites larger peak areas for chloroarsines were detected on PDMS-CAR or PDMS-CAR-DVB compared to PDMS 100. At most sites the volatile As species showed different sorption behavior on individual fibers during different sampling cam-paigns. For YNPHL01, (CH3)2AsS(CH3) was observed on PDMS-CAR during the first sampling campaign in June, but only on PDMS-CAR-DVB, not on PDMS-CAR, during the second sampling campaign in July. Another example is YNPNL04, where three volatile As species, (CH3)3As, (CH3)2AsCl and (CH3)2AsS(CH3), were found on PDMS 100 and PDMS-CAR-DVB in June. The PDMS-CAR fiber did not show any volatile As peaks. In September, only PDMS-CAR yielded one volatile As species, (CH3)2AsCl, while none was desorbed from the other two fibers.

Sampling duration reached from 18 to 98 hours with an average of 70 hours. In the laboratory, equilibrium in the ternary phase system sample solution – headspace – SPME fiber is achieved within minutes. Thus, shorter sampling times should be possible. Figure 76 shows a comparison of chromatograms sampling YNPRH05 for 94 hours and 2.3 hours (140 minutes) only. Most peaks are identical but the peak areas are smaller for the shorter sampling duration. The volatile As peak of (CH3)2AsCl at 3.8 minutes is missing in the chromatogram of the short-term sampling. Obvi-ously, equilibrium on the fiber takes longer in the field compared to laboratory experiments. On the other hand, a sampling duration of 70 hours might not be necessary considering that the two largest peaks for (CH3)2AsCl were detected at YNPRH05 and 06 after 23 hours of sampling. Another ex-planation could of course be that distinct species occur only sporadically, but as long as the sam-pling techniques are not optimized a natural variation cannot be verified.

Results 143

Figure 74 Volatile (CH3)2AsCl (first marked peak, mass spectrum above) and (CH3)AsCl2 (second marked peak, mass spectrum to the left) sampled with a PDMS-CAR-DVB fiber in July 2003 at YNPHL01

Figure 75 First discovery of volatile (CH3)2AsS(CH3) in a natural aquatic environment sampled with a PDMS-CAR fiber in June 2003 at YNPNL02

(CH3)2AsCl(CH3)2AsCl

(CH3)AsCl2

(CH3)AsCl2


Figure 76 Chromatograms from gas sampling with PDMS-CAR-DVB SPME fibers at YNPRH05 for 94 hours (black line) and 2.3 hours (gray line); while most peaks are identical, the (CH3)2AsCl peak (flagged in the chroma-togram) clearly only appears when longer sampling times are applied

The biggest challenge for a reliable interpretation is a reproducible sorption of all occurring volatile As species. As mentioned before, different sorption behavior was observed on individual fibers during different sampling campaigns. There seems to be a trend of fewer occurrence of (CH3)3As, (CH3)2AsCl, and (CH3)2AsS(CH3) from June and July to September and October. If this observed difference is not arbitrary it must be attributed to competing sorption reactions with other gases, maybe of seasonally changing composition. It is also striking that volatile As species were only detected in 23 of 57 gas samples even though the results of total volatile As from oxidation solu-tions confirm their presence in all samples. This contradiction might be explained by the presence of other volatile As species oxidized in NaOCl but not effectively sorbed on SPME fibers, such as (CH3)2AsH, (CH3)AsH2, or AsH3. Because of their low boiling points these compounds are not trapped on SPME fibers at ambient temperatures.

One more drawback is that under the GC-MS conditions described in Table 6 retention times still varied significantly between the different runs ( Table 12) requiring a MS-detector for unambiguous peak identification. Some of the chromatograms also showed a poor peak resolution from back-ground noise or a significant peak tailing over up to 2 minutes.

Table 12 Minimum, average and maximum retention times for the four volatile As species detected by GC-MS

retention time (CH3)3As (CH3)2AsCl (CH3)AsCl2 (CH3)2AsSCH3 min 1.42 1.88 5.32 4.72

mean 1.90 3.22 5.61 6.09 max 2.21 4.89 6.11 6.64

Results 145

Species-selective interpretation of other volatile elements detected besides volatile As was not an objective of this thesis. Thus, only speculations about the nature of these volatile elements can be made. The elements Li, Ba, B, Al, Si, S, and Cl can form simple volatile hydrides. For Cu, Fe, K, Sr, and Zn a transition to the gaseous phase is rather likely bonded to volatile organic substance, e.g., to carbonyls as reported for Mo and W (chapter 1). The chromatograms obtained were scanned for known mass spectra of any of these compounds, however, none could be identified. Detection of very low-molecular compounds such as BH3 (m = 13.8 g/mol) or H2S (m = 43 g/mol) as well as high-molecular compounds was excluded a priori by the selection of the mass scan range from 70-180 for an optimum resolution of the target compounds of volatile As.

4.3.5 Modeling volatile As speciation

In the WATEQ4F database no thermodynamic data for volatile As species were available. Thus, data for As(g), As2(g), As3(g), As4(g), AsCl3(g), AsF3(g), AsF5(g), AsH3(g), As4O6(g), AsS(g), and As4S4(g) compiled by Spycher and Reed (1989) were added to the PHREEQC input file ( Appendix 26). Partial pressures for all volatile As species considered were negligibly low ( Appendix 27). The highest partial pressure was found for As4S4 with 10-11 to 10-30 Vol%. The species AsH3 is third in abundance of gaseous As species with partial pressures as low as 10-23 to 10-34 Vol%. This insig-nificance of fractionation of the considered As species into the gas phase is in accordance with results from Spycher and Reed (1989) for geothermal fluids. The authors give various explanations for the discrepancy between detection of volatile As species in geothermal gases and the inability of modeling it. Possible reasons could be excess enthalpy boiling, leaching of previously deposited As minerals by dry steam or inaccurate stability constants for aqueous or gaseous species. The omission of volatile organometallic species in thermodynamic databases is acknowledged but con-sidered negligible. This conclusion was withdrawn from modeling Sb(CH3)3, one of the few vola-tile organometallics for which thermodynamic data were available and which yielded very low fugacities. Based on the results from As species-selective sampling described in the previous sec-tion ( 4.3.4), the insignificance of volatile organoarsines has to be questioned: Four different species were positively identified at various sites. Modeling of the results remained impossible since no thermodynamic data could be found for any of the species.

Still, the question remains about the nature of the volatile As species that were trapped in NaOCl solution but not sorbed on SPME fibers as discussed in the previous section ( 4.3.4). The explana-tion of Spycher and Reed (1989) that As is only carried in water droplets entrained in steam can be excluded from the use of water traps mounted before the oxidizing traps. Pokrovski et al. (2002) showed in their experimental study on As speciation in the vapor phase that with increasing tem-perature As(OH)3(g) becomes the predominant gaseous As species. Including As(OH)3(g) in the modeling by considering its Henry constant (log k = 11.2 at 25°C; Pokrovski et al. 2002) yielded partial pressures of 6.8 ⋅ 10-14 to 8.0 ⋅ 10-17 Vol%. Thus, As(OH)3(g) was the predominant species in 52 of 60 cases, but its partial pressures are still insignificantly low. Because of the lack of thermo-dynamic data for other volatile As species no further information could be obtained. A second


shortcoming is that microbial reactions are not considered in thermodynamic equilibrium modeling. The effect of this simplification is not known because there is so far no information about the im-portance of purely chemical volatilization compared to microbially catalyzed volatilization in geo-thermal waters.

4.3.6 Quantification of volatile metallics

In section 4.3.2 concentrations of volatile metallics were just given without further discussion. Be-cause quantification is not trivial, it shall be discussed in greater detail in the following. Table 13 and Table 14 show concentrations for volatile metallics from 55 gas samples in NaOCl trapping solutions and for volatile Cl from 38 gas samples in NaOH trapping solutions taken with the stan-dard setups 1 and 2 ( Figure 42). As explained in section 4.2.3.2, setup 1 and 2 integrate concentra-tions from underneath the gas hood directly above the sampled feature that enter the oxidizing bot-tle via one inlet and concentrations from ambient air approximately 25 cm above the sampled fea-ture that enter the oxidizing bottle via a second inlet. The comparisons made for the sampling sites YNPRH01, 04, and 22 between the setups 1 and 2 generally showed higher concentrations of vola-tile metallics, thus a more effective dissolution, when the tube was immersed in the oxidizing solu-tion (setup 1) compared to when it was left in the headspace above (setup 2). For both setups, gase-ous concentrations can only be estimated as upper limits for both compartments from two end-member models. Assuming that volatile metallics entered the oxidizing bottle exclusively via the short tube open to the atmosphere gives an upper limit for concentrations approximately 25 cm above the sampling site ( Table 13). If the concentrations of metallics in the oxidizing solution are exclusively attributed to diffusion via the long tube connected to the gas hood the data in Table 14 present an upper limit for the gaseous concentrations directly above the sampling site.

Table 13 Ambient concentrations [mg/m3] of volatile metallics for 55 NaOCl samples and of volatile Cl for 38 NaOH samples (setup 1 and 2, Figure 42) assuming that the volatile metallics all come from ambient air as ex-plained in section 4.2.3.2

mg/m3 As Al B Ba Cu Fe K Li SiO2 Sr Zn S Cl min 0.07 <12 <1.2 <0.09 <0.4 <0.8 <20 <0.12 <10 <0.03 <0.5 <155 78

mean 5.1 39 5.6 0.8 1.6 3.9 190 1.0 150 0.48 3.7 520 850 max 29 130 56 5.5 5.5 32 740 11 1200 1.4 42 4,300 4190

Table 14 Source concentrations [mg/m3] of volatile metallics for 55 NaOCl samples and of volatile Cl for 38 NaOH samples (setup 1 and 2, Figure 42) assuming that the volatile metallics all come from underneath the gas hood (section 4.2.3.2)

mg/m3 As Al B Ba Cu Fe K Li SiO2 Sr Zn S Cl min 0.9 <140 <14 <1.1 -4.2 <9.8 <240 <1.4 <123 <0.4 <5.6 <1,900 940

mean 62 460 67 9.5 19 46 2300 12.5 1850 5.8 45 6,230 10,240 max 350 1,580 670 66 66 390 8900 130 14300 16 510 52,000 50,300

The assumption of exclusive diffusion of volatile metallics from ambient air seems unlikely be-cause the distribution of volatile metallics is limited by diffusion, turbulent convection (eddy flux), dilution, hydrolysis, and oxidation in the atmosphere. To evaluate whether there is any significant

Results 147

amount of volatile metallics approximately 25 cm above hot springs at all, the results from setup 3 ( Figure 42) were used. Setup 3 was not connected to a gas hood and both valves were left open to the atmosphere. Volatile metallics were found in all of the 20 NaOCl samples and volatile Cl was found in all of the 10 NaOH samples. Accidental contamination was excluded by repeated labora-tory and field blank values. For 16 NaOCl and 8 NaOH samples setups 1 and 3 were run parallel. The results from setup 3 were used to determine the share of volatile metallics from ambient air in setup 1 and a corrected concentration of volatile metallics underneath the gas hood could be deter-mined as explained in section 4.2.3.2 ( Table 15). Only the 6 elements with the highest concentra-tions (As, Al, K, SiO2, Si and Cl) were considered. The average concentrations of all volatile met-allics decrease rapidly after degassing from a geothermal feature (assumed as 100%) to only about 6% in ambient air above the sampled feature. Table 15 shows that this decrease varies between 2% and 11% for the different sampling sites and between 3% and 10% for the elements considered.

Table 15 Concentrations of volatile metallics in ambient air (AA) approximately 25 cm above the sampling site compared to concentrations underneath a gas hood (GH) directly above the sampling site in mg/m3 (calculations from setup 1 and 3 ( Figure 42) as explained in section 4.2.3.2) and as percentage (AA:GH⋅100)

As (

AA

)

As (

GH

)

Al (

AA

)

Al (

GH

)

K (A

A)

K (G

H)

SiO

2 (A

A)

SiO

2 (G

H)

S (A

A)

S (G

H)

Cl (

AA

)

Cl (

GH

)

aver

age

% in

A

A p

er si

te

YNPGG02-4G 1.5* 1.3* 1.1 269 30 1185 109 5031 24 2342 96 1789 % in AA 118* 0.4 2.5 2.2 1.0 5.4 2.3 YNPGG03-4G 0.4 20 7.3 474 30 809 128 4063 24 3724 128 1565 % in AA 2.1 1.6 3.7 3.2 0.7 8.2 3.2 YNPGG04-4G 0.7 22 13 150 30 622 79 4664 23 280 112 1716 % in AA 3.3 8.5 4.8 1.7 8.3 6.5 5.5 YNPGG05-4G 0.5 4.5 8.7 155 70 552 1.1 339 25 304 94 1360 % in AA 10 5.6 13 0.3 8.3 6.9 7.3 YNPHL01-3G 0.01 1.9 4.5 54 3.0 18 121* 54* 101 1207 % in AA 0.4 8.3 17 222* 8.3 8.4 YNPNL02-2G 0.9 46 16 430 185 2432 2.4 2858 211 3315 % in AA 1.9 3.7 7.6 0.1 6.4 3.9 YNPRH01-3G 0.07 8.1 17 307 26 929 303 4367 811 10413 % in AA 0.8 5.4 2.8 6.9 7.8 4.7 YNPRH01-3G-tni 0.07* 0.07* 17 437 26 815 303 4918 811 10072 % in AA 94* 3.8 3.2 6.2 8.1 5.3 YNPRH04-3G 11.5* -38* 45 405 180 1268 482* 438* 100 15305 % in AA -30* 11 14 110* 0.7 8.6 YNPRH05-3G 9.2 76 2.7 262 54 1913 642 7068 431 11859 % in AA 12.1 1.1 2.8 9.1 3.6 5.7 YNPRH05-4G 10.6* 6.0* 24 543 217 3933 118 3068 161 5568 % in AA 177* 4.5 5.5 3.9 2.9 4.2 YNPRH05-4G-tni 10.6 77 24 396 217 3753 118 3450 161 2233 % in AA 14 6.2 5.8 3.4 7.2 7.3 YNPRH06-2G 7.0* 2.7* 22 341 115 2710 3.6 426 428 6103 % in AA 258* 6.6 4.3 0.8 7.0 4.7 YNPRH22-4G 4.5 147 45 1543 133 5220 3.1 2897 68 4638 205 1823 % in AA 3.1 2.9 2.5 0.1 1.5 11 3.6


Continuation of Table 15

As (

AA

)

As (

GH

)

Al (

AA

)

Al (

GH

)

K (A

A)

K (G

H)

SiO

2 (A

A)

SiO

2 (G

H)

S (A

A)

S (G

H)

Cl (

AA

)

Cl (

GH

)

aver

age

% in

A

A p

er si

te

YNPRH22-4G-tni 4.5 50 45 549 133 2501 3.1 37 68 818 205 786 % in AA 9.0 8.2 5.3 8.3 8.3 26 10.9 YNPRH23-4G 3.7 61 36 719 88 3538 2.9 1735 33 391 136 1989 % in AA 6.1 4.9 2.5 0.2 8.3 6.8 4.8 YNPRH23-4G-tni 136 1892 % in AA 7.2 7.2 average % in AA per element 5.7 5.2 6.1 3.3 5.5 9.8 5.8

* 7 samples for which the concentrations underneath the gas hood were lower than in ambient air; the reason for this uncommon distribution can either be natural caused by local turbulences leading to higher concentrations in ambient air than underneath the gas hood or (more likely) artificial caused by analytical errors at low gaseous concentrations; those values were not considered for the calculation of average percentage per site or per element in ambient air

If this decrease of 6% on average is also assumed for the other sites sampled with setup 1 corrected values can be calculated instead of the upper limits given in Table 13 and Table 14 according to the following equation:

DAtm

∆x∆c

∆x∆c

2

2

1

1⋅⋅

=+ with ∆c1 = 0.06 ⋅ ∆c2

DAtm

∆x∆x∆x∆c∆x∆c0.06

21

1222⋅⋅

=+

⋅

⋅⋅⋅

12

122 ∆xDAt

∆xm)∆x∆x(0.06∆c ⋅

⋅⋅⋅

⋅⋅=+

)∆x∆x(0.06∆x

DAt∆xm

∆c12

122 +⋅⋅⋅

⋅⋅

= with ∆x1 = 5 cm, ∆x2 = 60 cm

58.0DAt

∆xm∆c 2

2 ⋅⋅⋅

⋅= (compare calculation for setup 1 with ∆c2 = 0 in section 4.2.3.2)

From this assumption, the upper limits given in Table 14 were corrected by a factor of 0.58 to get an approximation for the concentrations underneath the gas hood. Concentrations given in section 4.3.2 are all concentrations underneath a gas hood corrected by this factor 0.58.

The results from the diffusion bowls (DBs) confirm the rapid decrease of volatile As with increas-ing distance from a geothermal feature. The DBs were set up about 1-2 m from the geothermal feature at the shores of Hazle Lake (YNPHL01, 03), around the hot springs at Frying Pan Spring (YNPNL02, 04), along “Milky Way” (YNPRH01), and around “Rainbow Growler” (YNPRH10). Concentrations of volatile metallics ranging from <0.1 to 36 µg/m3 are 2 to 3 orders of magnitude lower than in ambient air 25 cm above the feature (see digital database; Appendix 9). Figure 77 shows an example of decreasing concentrations of volatile As with increasing distance from the hot spring at “Rainbow Growler”.

Results 149

Figure 77 Setup of diffusion bowls around YNPRH10 (“Rain-bow Growler”); concentrations of volatile As decrease from 7.1 and 3.4 µg/m3 at DB04 and DB03 to 1.1 and 0.9 µg/m3 at DB05 and DB01; at DB06, the sampling site furthest away from the hot spring, concen-trations dropped below 0.1 µg/m3

The concentrations of volatile As discussed in the previous paragraphs with a minimum of about 500 µg/m3, an average of 36 mg/m3, and a maximum of 200 mg/m3 ( Table 14, corrected by the factor 0.58) directly above geothermal features are high compared to other values cited in literature. However, it also has to be seen that only few data exist for quantifying volatile As. A global bal-ance of Chilvers and Peterson (1987) estimates an average volatile As concentration over land of 2.8 ng/m3 in the northern hemisphere, and 1 ng/m3 in the southern hemisphere. Increased concen-trations of volatile As were quantified by Feldmann (1995) for waste heaps (up to 50 µg/m3) and digester gas (up to 25 µg/m3 for mesophilic digestion and up to 30 µg/m3 for thermophilic diges-tion). Wongsanoon et al. (1997) used NaOCl-trapping of volatile As from soils in a mining district in a similar way as it was done for the geothermal gases of Yellowstone within this research (sec-tion 2.7.5.1). Their concentrations in the NaOCl solution (13 µg As/L maximum, 1.5 µg As/L on average) are very similar to the ones found for Yellowstone gases (14 µg As/L maximum, 2 µg As/L on average) but Wongsanoon et al. (1997) did not try to convert the concentrations in trap-ping solution to concentrations in the gas phase and no data are given for exposure duration.

The concentrations derived from active pumping for 1 to 4.5 hours at four sampling sites are shown in Table 16. Compared to concentrations from passive sampling they are 1 to 2 orders of magnitude lower for volatile S, and 2 to 3 order of magnitude lower for all other volatile elements. With sev-eral tens to hundreds of µg/m3 they are similar to the concentrations determined by Feldmann (1995) for waste heaps and digester gas.

Table 16 Concentrations of volatile metallics [µg/m3] from active pumping for 1 to 4.5 hours

µg/m3 As Al B Ba Cu Fe K Li SiO2 Sr Zn S YNPGG04-2G-wGH-P 60 <600 <100 42 36 950 7,006 <10 <600 36 69 210,000 YNPHL01-2G-wGH-P 18 476 <50 19 37 516 4,014 <5 <300 17 <20 87,000 YNPRH01-2G-wGH-P 30 717 <100 26 183 146 7,834 <10 <600 29 <40 132,000 YNPRH06-2G-wGH-P 102 1,250 319 33 189 1,348 <800 <20 <1,200 17 281 351,000

DB 04DB 03

DB 06

DB 01 DB 05


Two processes may be responsible for the significant differences between active and passive gas sampling. Dilution with ambient air during pumping can not be excluded for active gas sampling because, with the simple Dräger hand pump, the gas flow rate could not be adjusted to a sampling volume equivalent to the amount of gas available at a specific site. Furthermore, active gas sam-pling was done in the afternoon, the time of lowest gas activity, whereas passive sampling over several days yielded an integrated signal of average gas concentrations. Concentrations from active sampling are thus likely to be too low. The high results from passive gas sampling on the other hand seem to conflict with the odor threshold of 1.6 mg/m3 for AsH3 at 25°C (NJ Hazardous Sub-stance Fact Sheets 1993). However, this odor threshold does not necessarily present an upper limit for volatile As concentrations. The absence of the typical garlic-like odor at all sampled features could also be explained by other volatile As species predominating, or masking by other gases such as H2S. Also, concentrations decrease rapidly already at 25 cm above the sampling site as shown before, and no olfactometric testing was done so close to the ground.

The preceding discussion shows how difficult it is to quantify volatile metallics. Nevertheless this study provides the first data on volatile metallics for the Yellowstone National Park and one of the largest database for quantification of volatile metallics in natural aquatic systems at all. The quanti-tative information is especially valuable in the light of toxicity. As described in section 2.5.3.1, a concentration of 160 µg/m3 can result in acute toxicity for exposure times over 1 hour (Office of Environmental Health Hazard Assessment 1999). However, the available data also showed that the significantly increased concentrations directly above a geothermal feature decrease rapidly to levels below detection limit within several decimeters to 1 m at maximum. While humans rarely approach geothermal features that close for longer exposure times, animals might be endangered because of the fact that they breathe air from lower distances above ground and that, especially in winter times, they assemble around the warm geothermal features ( Figure 78).

Figure 78 Bison herd graz-ing at a hot spring – a potential health hazard considering in-creased concentrations of vola-tile metallics, especially As, in the vicinity of geothermal fea-tures

The observed decrease in life expectancy of elk populations in areas with a large number of geo-thermal features compared to control populations is so far attributed to high As concentrations in

Summary of chapter 4 151

forage species found in aquatic and riparian habitats and to accelerated dental aging caused by high F concentrations (Garrott et al. 2003; Kocar et al. 2004). A chronic effect of high volatile As con-centrations was not considered so far. An acute toxic effect of volatile As species seems even less likely. However, several deaths of animals were reported in the park´s history attributed to accumu-lation of toxic gases denser than air in topographically low areas especially during inverted atmos-pheric conditions (Fournier 1959; Frisbee 1960; Jagger 1899; Love and Good 1970; Weed 1889). The most recent death of 5 bisons occurred around March, 1st, 2004 just north of the Norris Geyser Basin along the Gibbon River downstream of multiple gas vents. Current expert opinion is that high concentrations of H2S (up to > 200 ppm) and CO2 asphyxiated the bisons (Heasler and Ja-worowski 2004). Again, volatile As - also more dense than air and with an even higher toxicity potential than CO2 or H2S - is so far no object of research.


As an example for an aqueous environment predestined for the release of volatile metallics Yellow-stone National Park was investigated. It is located over the largest continental hot spot world wide and contains more than 10,000 geothermal features. After the first detection of volatile As during a reconnaissance study in 2002, 5 study areas (Nymph Lake, Hazle Lake, Ragged Hills, Gibbon Gey-ser Basin, and Lower Geyser Basin) were sampled for their water and gas chemistry. Steam-heated sulfate-dominated waters with low pH and As concentrations between 20 and 2,000 µg/L were differentiated from deep thermal chloride-dominated waters with near neutral pH and As concen-trations between 1,050 and 11,000 µg/L. The precipitation of As minerals was thermodynamically calculated for Hazle Lake (orpiment supersaturation) and actually observed for a low temperature hot spring at Ragged Hills (precipitation of an amorphous Fe, S, and As rich material).

Volatile As was found in all samples together with volatile Al, B, Ba, Cu, Fe, K, Li, Si, Sr, Zn, and S. By species-selective sampling on SPME fibers, four volatile As species were detected: (CH3)3As, (CH3)AsCl2, (CH3)2AsCl, and (CH3)2AsSCH3. The latter three were all found for the first time in a natural environment. Modeling of the gaseous As species proved impossible because of a lack of thermodynamic constants for the species detected. The volatile inorganic As, As-S, As-F, and As-Cl species for which thermodynamic data were available showed negligible partial pres-sures of <10-11 Vol%. Total concentrations in the trapping solution from passive, diffusion based sampling were converted to concentrations in gas phase using Fick´s first law. Medium concentra-tions were 8 and 28 mg/m3 for the two clusters within steam-heated waters and 11 and 89 mg/m3 for the two clusters within deep thermal waters. Concentrations from active gas sampling by pump-ing were lower (20-100 µg/m3) but are supposed to be subject to significant dilution by withdraw-ing ambient air from over-pumping. The detected concentrations are significantly increased com-pared to the acute toxicity limit for 1 hour exposure (160 µg/m3) with respect to AsH3 that, how-ever, is probably not the predominant volatile species. Even though concentrations decrease rapidly with increasing distance from a geothermal feature (to about 6% the original value in 25 cm above


the sampled structures and 0.06-0.006% in 1 to 2 m horizontal distance) animals grazing and warming up at hot springs especially during winter times might be exposed to significant amounts of volatile metallics, including As. Both chronic effects such as the observed decrease in life expec-tancy of elk populations in areas with a large number of geothermal features compared to control populations as well as acute poisoning such as the asphyxiation of bisons north of Norris Geyser Basin in March 2004 could be attributed to high volatile As concentrations but have so far never been investigated.

5 RECOMMENDATIONS

During the present research, a robust method was developed for in-situ collection of volatile metal-lics from aqueous environments. Volatile As was detected in significant concentrations in geother-mal waters of Yellowstone National Park and species-selective sampling showed three new species never detected in natural waters or gases before. Yet, this is a mere drop in the ocean.

Sampling and quantification of total volatile metallic concentrations must be optimized. An alterna-tive to passive diffusion based sampling could be active gas sampling by low-flow pumping avoid-ing the withdrawal of ambient air by over-pumping and the dilution of target concentrations. Quan-tification of individual As species would be most desirable, but before that basic information must be obtained about preferential sorption of As species on different sorption material and about pos-sible competitive sorption with gases that could typically occur in environments where volatile metallics are found, such as CO2, CH4, H2S, and SO2. For an economic screening, sampling dura-tion must be minimized. A mobile GC-MS equipment set up directly in the field would help to avoid storage losses or species conversions. Cryotrapping could be applied to detect lower concen-trations and provide sample stability, yet the problem of transporting and handling liquid N2 in the field has to be considered. For analysis, the GC program must be further optimized to achieve ex-actly reproducible retention times. Mass spectra of other volatile metallics so far missing in stan-dard databases must be gathered from recent publications or obtained by own experiments.

Yellowstone National Park provides a wide variety of hydrogeochemical conditions and further correlation between hydrogeochemical parameters and volatile metallics, especially As, will surely be found in more detailed studies. One of the most interesting questions is about the origin of the volatile As. Is its formation purely chemical and if so does it correlate with geothermal activity? Do microbes contribute significantly to volatilization and which of the various microorganisms living under extreme conditions of pH and heat control the volatilization of each species? Are there sea-sonal variations in volatilization rates? Another interesting question is about the stability and distri-bution of the volatile metallics. Samples must be taken in vertical and horizontal sections from the source to determine the rate of hydrolysis and oxidation of the volatile species. Site-specific mete-orological data will be needed. Laboratory experiments can help to obtain hydrolysis and oxidation rates, but they are mostly limited to one-metal-standard solutions not taking into account the pres-ence and effects of other gases. The stability and distribution of volatile metallics will directly af-fect studies about chronic and acute poisoning of humans and animals by volatile metallics. Further interdisciplinary collaboration of geologists, hydrogeologists, chemists, microbiologists, biologists, and toxicologists is required to achieve more detailed knowledge.

6 REFERENCES Abedin JMD, Cresser MS, Meharg A, Feldmann J, Cotter-Howells J (2002) Arsenic Accumulation and Metabolism in

Rice (Oryza sativa L.). Environ Sci Technol 36, 962-968. Ackely KL, Hymer CB, Sutton JA, Caruso J (1999) Speciation of arsenic in fish tissue using microwave-assisted extrac-

tion followed by HPLC-ICP-MS. J Anal Atom Spectrom 14, 845-850. Agterdenbos J, Bax D (1986) Mechanisms in hydride generation AAS. Fresen Z Anal Chem 323 (7), 783-787. Ahmann D, Krumholz LR, Hemond HF, Lovley DR, Morel FMM (1997) Microbial mobilization of arsenic from sedi-

ments of the Aberjona watershed. Environ Sci Technol 31 (10), 2923-2930. Allen ET, Day AL (1935) Hot springs of the Yellowstone National Park. Carnegie Inst Wash Publ 466, 525. Amouroux D, Donard OFX (1996) Maritime emission of selenium to the atmosphere in eastern Mediterranean seas.

Geophys Res Lett 23 (14), 1777-1780. Amouroux D, Tessier E, Pécheyran C, Donard OFX (1998) Sampling and probing volatile metal(loid) species in natural

waters by in-situ purge and cryogenic trapping followed by gas chromatography and inductively coupled plasma mass spectrometry (P-CT-GC-ICP/MS). Anal Chim Acta 377 (2-3), 214-254.

Anderson LCD, Bruland KW (1991) Biogeochemistry of arsenic in natural waters: the importance of methylated species. Environ Sci Technol 25 (3), 420-427.

Andino JM, Butler JW (1991) A study of the stability of methanol-fueled vehicle emissions in Tedlar bags. Environ Sci Technol 25, 1644-1646.

Andreae MO (1978) Distribution and speciation of arsenic in natural waters and some marine algae. Deep Sea Res 25, 391-402.

Andreae MO (1979) Arsenic speciation in seawater and industrial waters. Limnol Oceanogr 24, 440-452. Andreae MO, Asmodé JF, Foster P, Van´t Dack L (1981) Determination of antimony(III), antimony(V), and methylanti-

mony species in natural waters by atomic absorption spectrometry with hydride generation. Anal Chem 53, 1766-1771.

Andreae MO, Froehlich PN (1984) Arsenic, antimony, and germanium biogeochemistry in the Baltic Sea. Tellus 36, 101-117.

Andrewes P, Cullen WR, Polishchuk E (2000) Arsenic and antimony biomethylation by Scopulariopsis brevicaulis: Interaction of arsenic and antimony compounds. Environ Sci Technol 34 (11), 2249-2253.

Aposhian HV (1997) Enzymatic methylation of arsenic compounds: II. An overview. In Abernathy C, Calderon R, Chappell W (eds) Human Health II, Chapman and Hall, London, 296-321.

Aposhian HV, Zakharyan RA, Wildfang EK, Healy SM, Gailer J, Radabaugh TR, Bogdan GM, Powell LA, Aposhian MM (1999) How is inorganic arsenic detoxified?. in Chappell WR, Abernathy CO, Calderon RL (eds) Arsenic expo-sure and health effects. Elsevier, 289-297.

Aposhian HV, Gurzau ES, Le XC, Gurzau A, Healy SM, Lu X, Ma M, Yip L, Zakharyan RA, Maiorino RM, Dart RC, Tircus MG, Gonzalez-Ramirez D, Morgan DL, Avram D, Aposhian MM (2000) Occurrence of monomethylarson-ous acid in urine of humans exposed to inorganic arsenic. Chem Res Toxicol 13 (8), 693-697.

Arcand GM (1957) Distribution of tripositive arsenic between hydrochloric acid solutions and β-β’-dichlorodiethyl ether. J Am Chem Soc 79 (8), 1865-1870.

Armienta MA, Rodriguez R, Aguayo A, Ceniceros N, Villasenor G, Cruz O (1997) Arsenic contamination of groundwa-ter at Zimapán, Mexico. Hydrogeol J 5, 39-46.

Armienta MA, Villasenor G, Rodriguez R, Ongley LK, Mango H (2001) The role of arsenic-bearing rocks in groundwa-ter pollution at Zimapán Valley,México. Environ Geol 40 (4-5), 571-581.

Arnórsson S, Andrésdóttir A, Gunnarsson I, Stefánsson A (1998) New calibration for the quartz and Na/K geothermome-ters – valid in the range 0-350°C (in Icelandic). Proc Geoscience Society of Iceland Annual Meeting, April 1994, 42-43.

Arnórsson S (ed) (2000) Isotopic and chemical techniques in geothermal exploration, development and use – Sampling methods, data handling, interpretation. International Atomic Energy Agency, Austria.

Arthur CL, Pawliszyn J (1990) Solid phase microextraction with thermal desorption using fused silica optical fibers. Anal Chem 62 (19), 2145-2148.

Ashe AJ, Ludwig EG (1986) The exchange reaction of tetramethyldipnictogens with dimethyldichalcogenides. Or-ganomet Chem 308, 289-296.

Aström O (1982) Flow injection analysis for the determination of bismuth by atomic absorption spectrometry with hy-dride generation. Anal Chem 54, 190-193.

Azcue JM, Nriagu JO (1994) Arsenic: historical perspectives. In.Nriagu JO (ed) Arsenic in the environment – Part I: Cycling and characterization. Adv Environ Sci 26, 1-15.

Azcue JM, Murdoch A, Rosa F, Hall GEM (1994) Effects of abandoned gold mine tailings on the arsenic concentrations in water and sediments of Jack of Clubs Lake, BC. Environ Technol 15, 669–678.

Baes CF, Mesmer RE (1976) The hydrolysis of cations. John Wiley, New York. Ball JW, Nordstrom DK (1991) User's manual for WATEQ4F, with revised thermodynamic data base and test cases for

calculating speciation of major, trace, and redox elements in natural waters. US Geol Survey Open-File Report 91-183.

156 References

Ball JW, Nordstrom DK, Jenne EA, Vivit DV (1998a) Chemical analyses of hot springs, pools, geysers, and surface waters from Yellowstone National Park, Wyoming, and vicinity, 1974-1975. US Geol Survey Open-File Report 98-182.

Ball JW, Nordstrom DK, Cunningham KM, Schoonen MAA, Xu Y, DeMonge JM (1998b) Water-chemistry and on-site sulfur-speciation data for selected springs in Yellowstone National Park, Wyoming, 1994-1995. US Geol Survey Open-File Report 98-574.

Ball JW, Nordstrom DK (2001) Application of D/H and 18O/16O values to the interpretation of subsurface fluid mixing at Yellowstone National Park, USA. Applied Isotope Geochemistry IV, June 25-29 2001, Pacific Grove, California.

Ball JW, Nordstrom DK, McCleskey RB, Schoonen MAA, Xu Y (2001) Water-chemistry and on-site sulfur-speciation data for selected springs in Yellowstone National Park, Wyoming, 1996-1998. US Geol Survey Open-File Report 01-49.

Ball JW, McCleskey RB, Nordstrom DK, Holloway JM, Verplanck PL (2002) Water-chemistry data for selected springs, geysers, and streams in Yellowstone National Park, Wyoming, 1999-2000. US Geol Survey Open-File Report 02-382.

Ballantyne JM, Moore JN (1988) Arsenic geochemistry in geothermal systems. Geochim Cosmochim Acta 52, 475–483. Baltussen E, Sandra P, David F, Cramers J (1999) Stir bar sorptive extraction (SBSE), a novel extraction technique for

aqueous samples: theory and principles. J Microcol Sep 11 (10), 737-747. Bargar KE, Fournier RO (1988) Effects of glacial ice on subsurface temperatures of hydrothermal systems in Yellow-

stone National Park, Wyoming; Fluid inclusion evidence. Geology 16, 1077-1080. Becker J (2004) High resolution aerial and ground mapping of geothermal features in Ragged Hills, Yellowstone Na-

tional Park. Unpublished master thesis, Technische Universität Bergakademie Freiberg, Department of Geology. Bednar AJ, Garbarino JR, Ranville JF, Wildeman TR (2002) Preserving the distribution of inorganic arsenic species in

groundwater and acid mine drainage samples. Environ Sci Technol 36, 2213-2218. Bednar AJ, Garbarino JR, Burkhardt MR, Ranville JF, Wildeman TR (2004) Field and laboratory arsenic speciation

methods and their application to natural-water analysis. Water Res 38, 355-364. Bentley R, Chasteen TG (2002a) Arsenic curiosa and humanity. Chem Educator 7 (2), 51-60. Bentley R, Chasteen TG (2002b) Microbial methylation of metalloids: arsenic, antimony, and bismuth. Microbiol Mol

Biol R 66 (2), 250-271. Berg M, Tran HC, Nguyen TC, Pham HV, Schertenleib R, Giger W (2001) Arsenic contamination of groundwater and

drinking water in Vietnam: a human health threat. Environ Sci Technol 35, 2621–2626. Bergmann A, Glindemann D (1997) Analytik flüchtiger Metalloid- und Metall-Verbindungen für die Untersuchung ihrer

Wurzelraum-Atmosphäre-Austauschprozesse. Wissenschaftlicher Sachbericht, Förderkennzeichen UFZ-21/97; Um-weltforschungszentrum Leipzig-Halle GmbH.

Bermejo-Barrera P, Moreda-Pineiro J, Moreda-Pineiro A, Bermejo-Barrera A (1998) Selective medium reactions for the arsenic(III), arsenic (V), dimethylarsonic acid and monomethylarsonic acid determination in waters by hydride gen-eration on-line electrothermal atomic absorption spectrometry with in situ preconcentration on Zr-coated graphite tubes. Anal Chim Acta 374, 231-240.

Biggar JW, Jayaweera GR (1993) Measurement of selenium volatilization in the field. Soil Sci 155 (1), 31-36. Biginelli P (1900a) Composizione e constituzione chimica del gas arsenical delle tappezzerie. Nota I Atti Reala Accad

Lincei 9, 210-214. Biginelli P (1900b) Composizione e constituzione chimica del gas arsenical delle tappezzerie. Nota II Atti Reala Accad

Lincei 9, 242-249. Bissen M, Frimmel FH (2000) Speciation of As(III), As(V), MMA and DMA in contaminated soil extracts by HPLC-

ICP/MS. Fresen J Anal Chem 367, 51-55. Blasco F, Jeanjean R, Gaudin C (1972) Absorption of arsenate ions by Chlorellas - Partial reduction of arsenate to ar-

senite. Characteristics Comptes Rendus des Seances de l'Academie des Sciences, Serie D: Sciences Naturelles, 275 (12), 1223-1226.

Bothe JV, Brown PW (1999) Arsenic immobilization by calcium arsenate formation. Environ Sci Technol 33 (21), 3806-3811.

Borho M, Wilderer P (1997) A reliable method for preservation and determination of arsenate(III) concentrations in groundwater and water works samples. J Water SRT Aqua 46, 138-143.

Braman RS, Foreback CC (1973) Methylated forms of arsenic in the environment. Science 182, 1247-1249. Branfireun BA, Heyes A, Roulet NT (1996) The hydrology and methylmercury dynamics of a Precambrian Shield head-

water peatland. Water Resour Res 32 (6), 1785-1794. Bright DA, Dodd M, Reimer KJ (1996) Arsenic in subarctic lakes influenced by gold mine effluent: the occurrence of

organoarsenicals and 'hidden' arsenic. Sci Total Environ 180 (2), 165-182. British Geological Survey Department of Public Health Engineering (2001) Arsenic contamination of groundwater in

Bangladesh. In: Kinniburgh DG, Smedley PL (eds) British Geological Survey (Technical Report, WC/00/19. 4 Vol-umes). British Geological Survey, Keyworth.

Brookins DG (1988) EH-pH diagrams for geochemistry. Springer, Berlin. Brooks WE (2004) Arsenic statistics and information. US Geol Survey, Mineral Commodity Summaries,

http://minerals.usgs.gov/minerals/pubs/commodity/arsenic/ (27.03.2003), 24-25. Brünnich JC (1909) Rep Aust Ass Adv Sci 12, 129.

References 157

Brumbaugh WG, Ingersoll CG, Kemble NE, May TW, Zajicek JL (1994) Chemical characterization of sediments and pore water from the upper Clark Fork river and Milltown Reservoir, Montana. Environ Toxicol Chem 13 (12), 1971-1983.

Buchet JP, Lauwerys R, Roels H (1981) Comparison of the urinary excretion of arsenic metabolites after a single oral dose of sodium arsenite, monomethylarsonate, or dimethylarsinate in man. Int Arch Occup Environ Health 48, 71-79.

Bundesminister für Ernährung, Landwirtschaft und Forsten (1974) Verordnung zur Neufassung der Verordnung über Anwendungsverbote und –beschränkungen für Pflanzenschutzmittel vom 31. Mai 1974. Bundesgesetzbl 1974/I:1204-1210, Bonn.

Bzezinska A, Loon van J, Williams D, Oguma K, Fuwa K, Haraguchi IH (1983) A study of the determination of di-methylmercury and methylmercury chloride in air. Spectrochim Acta 38B (10), 1339-1346.

Cabon JY, Cabon N (2000) Speciation of major arsenic species in seawater by flow injection hydride generation atomic absorption spectrometry. Fresen J Anal Chem 368, 484-489.

Cai Y, Jaffe R, Jones R (1997) Ethylmercury in the soils and sediments of the Florida Everglades. Environ Sci Technol 31, 302-305.

Cai Y, Jaffe R, Jones R (1999) Interactions between dissolved organic carbon and mercury species in surface waters of the Florida Everglades. Appl Geochem 14, 395-407.

Calogirou A, Richter Larsen B, Brussol C, Duane M, Kotzias D (1996) Decomposition of terpenes by ozone during sampling on Tenax. Anal Chem 68(9), 1499-1506.

Cantoni GL (1953) S-Adenosylmethionine; A new intermediate formed enzymatically from L-methionine and adenosi-netriphosphate. J Biol Chem 204, 403-416.

Cao XL, Hewitt CN (1994) Build-up of artifacts on adsorbents during storage and its effect on passive sampling and gas chromatography - flame ionization detection of low concentrations of volatile organic compounds in air. J Chroma-togr A 688, 368-374.

Carrero P, Malave A, Burguera JL, Burguera M, Rondon C (2001) Determination of various arsenic species by flow injection hydride generation atomic absorption spectrometry: investigation of the effects of the acid concentration of different reaction media on the generation of arsines. Anal Chim Acta 438, 195-204.

Carson BL, Ellis HV, McCann JL (1986) Toxicology and biological monitoring of metals in humans - including feasibil-ity and need. Lewis Publishers. Inc Chelsea.

Challenger F, Higginbottom C, Ellis L (1933) Formation of organometalloid compounds by microorganisms. I. Trimethylarsine and dimethylarsine. Chem Soc 95-101.

Challenger F (1945) Biological methylation. Chem Rev 36, 315-361. Challenger F (1951) Biological methylation. Adv Enzymol 12, 429-491. Challenger F, Lisle DB, Dransfield PB (1954) Studies in biological methylation. Part XIV. The formation of trimethylar-

sine and dimethyl selenide in mould cultures from methyl sources containing 14C. J Chem Soc, 1760-1771. Chausseau M, Roussel C, Gilon N, Mermet JM (2000) Optimization of HPLC-ICP-AES for the determination of arsenic

species. Fresen J Anal Chem 366, 476-480. Chen SL, Dzeng SR, Yang MH, Chlu KH, Shieh GM, Wal CM (1994) Arsenic species in groundwaters of the Blackfoot

disease areas, Taiwan. Environ Sci Technol 28, 877–881. Cheng CN, Focht DD (1979) Production of arsine and methylarsine in soil and in culture. Appl Environ Microbiol 38,

494-498. Cherry JA, Shaikh AU, Tallmann DE, Nicholson RV (1979) Arsenic species as an indicator of redox conditions in

groundwater. J Hydrol 43 (1-4), 373-392. Chilvers DC, Peterson PJ (1987) Global cycling of arsenic. In: Hutchinson TC, Meema KM (eds) Lead, mercury, cad-

mium, and arsenic in the environment. SCOPE 31, 279-301, John Wiley and Sons. Chiu VQ, Hering JG (2000) Arsenic adsorption and oxidation at manganite surfaces. 1- Method for simultaneous deter-

mination of adsorbed and dissolved arsenic species. Environ Sci Technol 34 (10), 2029-2034. Chou S, Odin M, Sage GW, Little S (2000) Toxicological profile for arsenic. US Department of Health and Human

Services, Public Health Service, Agency for Toxic Substances and Disease Registry. Christiansen RL (2001) The Quaternary and Pliocene Yellowstone Plateau Volcanic Field of Wyoming, Idaho, and Mon-

tana. US Geol Survey Professional Paper 729-G. Christiansen RL, Foulger GR, Evans JR (2002) Upper-mantle origin of the Yellowstone hotspot. Geol Soc Am Bull 114

(10), 1245-1256. Chukhlantsev VG (1956) Solubility products of arsenates. J Inorg Chem (USSR) 1, 1975-1982. Church PE, Vroblesky DA, Lyford FP, Willey RE (2002) Guidance on the use of passive-vapor-diffusion samplers to

detect volatile organic compounds in ground-water-discharge areas, and example applications in New England. US Geol Survey Water-Resour Invest Rept 02-4186.

Cicerone RJ, Oremland RS (1988) Biogeochemical aspects of atmospheric methane. Global Biogeochem Cy 2, 299-327. Clayton GD, Clayton FE (eds) (1981) Patty's industrial hygiene and toxicology. Vol 2 A-C: Toxicology. 3rd ed, John

Wiley. Clemens JB, Lewis RG (1988) In: Keith LH (ed) Principles of Environmental Sampling, ACS Professional Reference

Book, 287.

158 References

Cleverley JS, Benning LG, Mountain BW (2003) Reaction path modelling in the As-S system: a case study for geother-mal As transport. Appl Geochem 18, 1325-1345.

Col M, Col C, Soran A, Sayli BS, Ozturk S (1999) Arsenic related Bowen's disease, Palmar keratosis, and skin cancer. Environ Health Persp 107, 687-689.

Conrad R (1996) Soil Microorganisms as Controllers of Atmospheric Trace Gases (H2, CO, CH4, OCS, N2O, and NO). Microbiol Rev 60(4), 609-640.

Coury L (1999) Conductance measurements, Part 1: Theory. Current Separations.com, 18:3 http://www.currentseparations.com/issues/18-3/cs18-3c.pdf.

Cox DP, Alexander M (1973) The production of trimethylarsine gas from various arsenic compounds by three sewage fungi. Bull Environ Contam Toxicol 9, 84-88.

Craig H, Boato G, White DE (1956) Isotope geochemistry of thermal waters. National Research Council Publication 400, 29-38.

Craig PJ (1986) Organometallic compounds in the environment. Longman, Harlow, 368. Craig PJ (1989) In: Hartley FR (ed) Chemistry of metal-carbon bond. Vol.5, Wiley, Chicester, 437. Crecelius EA, Robertson DE, Fruchter JS, Ludwick JD (1976) Chemical forms of mercury and arsenic emitted by a

geothermal power plant. In: Hemphil DD (ed) Trace Substances in Environmental Health, University of Missouri, Columbia, 287-293.

Crecelius EA (1977a) Arsenite and arsenate levels in wine. Bull Environ Contam Toxicol 1, 227-230. Crecelius EA (1977b) Changes in the chemical speciation of arsenic following ingestion by man. Environ Health Persp

19, 147-150. Crecelius EA, Bloom NS, Cowan CE, Jenne EA (1986) Speciation of selenium and arsenic in natural waters and sedi-

ments: arsenic speciation. Electric Power Research Institute Vol.2, EA-4641, Project 2020-2. Creed JT, Magnuson ML, Brockhoff CA (1996) Arsenic determination in saline waters utilizing a tubular membrane as a

gas-liquid separator for hydride generation inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 11, 505-509.

Crompton TR (1998) Occurrence and analysis of organometallic compounds in the environment. Wiley, New York, 237. Cullen WR, McBride BC, Reglinski J (1984) The reduction of trimethylarsine oxide to trimethylarsine by thiols: a

mechanistic model for the biological reduction of arsenical. J Inorg Biochem 21, 45-60. Cullen WR and Dodd (1989) Arsenic speciation in clams of British Columbia. Appl Organomet Chem 3(1), 79-88. Cullen WR, McBride BC, Manji H, Pickett AW, Reglinski J (1989) The metabolism of methylarsine oxide and sulfide.

Appl Organomet Chem 3, 71-78. Cullen WR, Reimer KJ (1989) Arsenic speciation in the environment. Chem Rev 89, 713-764. Cullen WR, Li H, Pergantis SA, Eigendorf GK, Mosi AA (1995) Arsenic biomethylation by the microorganism

Apiotrichum humicola in the presence of L-methionine-methyl-d3. Appl Organomet Chem 9 (7), 507-515. Daus B, Mattusch J, Wennrich R, Weiss H (2002) Investigation on stability and preservation of arsenic species in iron

rich water samples. Talanta 58, 57-65. Davidson PJ, Lappert MF, Pearce R (1976) Metal σ-hydrocarbyls, MRn, stoichiometry, structures, stabilities, and thermal

decomposition pathways. Chem Rev 76, 219-242. De Bettencourt AMM, Andreae MO (1991) Refractory arsenic species in estuarine waters. Appl Organomet Chem 5,

111-116. De Bettencourt AMM, Florencio MH, Duarte MFN, Gomes MLR, Vilas Boas LFC (1994) Refractory methylated arsenic

compounds in estuarine waters: Tracing back elusive species. Appl Organomet Chem 8, 43-56. Delnomdedieu M, Basti MM, Otvos JD, Thomas DJ (1994) Reduction and binding of arsenate and dimethylarsinate by

glutathione: a magnetic resonance study. Chem Biol Interact 90, 139-155. Del Razo LM, Arellano MA, Cebrián ME (1990) The oxidation states of arsenic in well-water from a chronic arsenicism

area of northern Mexico. Environ Pollut 64, 143–153. Del Razo LM, García-Vargas GG, Hernández MC, Gómez-Munoz A, Cebrián ME (1999) Profile of urinary arsenic

metabolites in children chronically exposed to inorganic arsenic in Mexico. In: Chappell WR, Abernathy CO, Calderon RL (eds) Arsenic exposure and health effects, Elsevier, 281-287.

Department of Energy (2003) News Release October 21, 2003, http://newsdesk.inel.gov/contextnews.cfm?ID=472. Devesa V, Martínez A, Súñer MA, Benito V, Vélez D, Montoro R (2001) Kinetic study of transformations of arsenic

species during heat treatment. J Agric Food Chem 49, 2267-2271. Doak GO, Freedman LD (1970) Organometallic compounds of arsenic, antimony, and bismuth John Wiley, New York. Dobson PF, Kneafsey TJ, Hulen J, Simmons A (2003) Porosity, permeability, and fluid flow in the Yellowstone geo-

thermal system, Wyoming. J Volcanol Geotherm Res 123, 313-324. Donahoe-Christiansen J, D'Imperio S, Jackson CR, Inskeep WP, McDermott TR (2004) Arsenite-oxidizing hydrogeno-

baculum strain isolated from an acid-sulfate-chloride geothermal spring in Yellowstone National Park. Appl Environ Microbiol 70, 1865-1868.

Donard OFX and Weber JH (1988) Volatilization of tin as stannane in anoxic environment. Nature 332, 339-341. Duncan RA (1982) A captured island chain in the Coast Range of Oregon and Washington. J Geophys Res 87, 10827-

10837. Dzurisin D, Yamashita KM (1987) Vertical surface displacements at Yellowstone caldera, Wyoming, 1976-1986. J Geo-

phys Res 92, 13753-13766.

References 159

Dzurisin D, Savage JC, Fournier RO (1990) Recent crustal subsidence at Yellowstone caldera. B Volcanol 52, 247-270. Dzurisin D, Yamashita K, Kleinmann JW (1994) Mechanisms of crustal uplift and subsidence at the Yellowstone caldera.

B Volcanol 56, 261-270. Dzurisin D, Wicks C, Thatcher W (1999) Renewed uplift at the Yellowstone Caldera measure by leveling surveys and

satellite radar interferometry. B Volcanol 61, 349-355. Ebdon L, Hill S, Ward RW (1987) Directly coupled chromatography-atomic spectroscopy. Part 1. Directly coupled gas

chromatography-atomic spectroscopy. A review. Analyst 111, 1113-1138. Edmonds JS, Francesconi KA, Cannon JR, Raston CL, Skelton BW, White AH (1977) Isolation, crystal structure and

synthesis of arsenobetaine, the arsenical constituent of the western rock lobster Panulirus longipes cygnus George. Tetrahedron Lett 18, 1543-1546.

Edmonds JS, Francesconi KA (1981) Arseno-sugars from brown kelp (Ecklonia radiata) as intermediates in cycling of arsenic in a marine ecosystem. Nature 289, 602-604.

Edmonds JS, Francesconi KA, Healy PC, White AH (1982) Isolation and crystal structure of an arsenic-containing sugar sulphate from the kidney of the giant clam, Tridacna maxima. X-ray crystal structure of (2S)-3- [5-deoxy-5-(dimethylarsinoyl)- B-D-ribofuranosyloxy] -2-hydroxypropyl hydrogen sulphate. Journal of the Chemical Society, Perkin Transactions I, 2989-2993.

Edmonds JS, Shibata Y, Francesconi KA, Rippingale RJ, Morita M (1997) Arsenic transformations in short marine food chains studies by HPLC-ICPMS. Appl Organomet Chem 11, 281-287.

Edwards M, Patel S, McNeill L, Chim H, Frey M, Eaton AD, Antweller RC, Taylor HE (1998) Considerations in As analysis and speciation. J Am Water Works Assoc 90, 103-113.

Ehrlich HL (2002) Geomicrobiology, 4th ed., Marcel Dekker Inc, New York. Ellenhorn MJ (1997) Medical toxicology, diagnosis and treatment of human poisoning 2nd ed. Williams and Wilkins Co. Ellis A, Mahon W (1964) Natural hydrothermal system and experimental hot-water/rock interactions. Geochim Cosmo-

chim Acta 28, 1323-1357. Elsenbroich C, Salzer A (1992) Organometallics – A concise introduction, VCH, Weinheim. Environmental Protection Agency (2001) 40 CFR Parts 9, 141, and 142 National Primary Drinking Water Regulations;

Arsenic and Clarifications to Compliance and New Source Contaminants Monitoring; Final Rule. Part VIII. Federal Register 66 (14), 6976-7066.

Essington ME (1988) Division S-2 - soil chemistry: solubility of barium arsenate. Soil Sci Soc Amer J 52, 1566-1570. Ewers GR, Keays RR (1977) Volatile and precious metal zoning in the Broadlands geothermal field, New Zealand. Econ

Geol 72, 1337-1354. Fahlquist L, Janik CJ (1992) Procedures for collecting and analyzing gas samples from geothermal systems. US Geol

Survey Open-File Report 92-211. Falter R, Hintelmann H, Quevauviller P (1999) Conclusion of the workshop on sources of error in methylmercury deter-

mination during sample preparation, derivatisation and detection. Chemosphere 39, 1039-1049. Fanning JC (2000) The chemical reduction of nitrate in aqueous solution. Coord Chem Rev 199, 159-179. Farago ME, Kavanagh P (1999) Proportions of arsenic species in human urine. In: Chappell W.R., Abernathy C.O.,

Calderon R.L. (eds): Arsenic exposure and health effects 325-334 Elsevier. Feeney R, Kounaves SP (2000) On-site analysis of arsenic in groundwater using a microfabricated gold ultramicroelec-

trode array. Anal Chem 72, 2222-2228. Feldmann J, Grümping R, Hirner AV (1994) Determination of volatile metal and metalloid compounds in gases from

domestic waste deposits with GC/ ICP-MS. Fresen J Anal Chem 350, 228-234. Feldmann J (1995) Erfassung flüchtiger Metall- und Metalloidverbindungen in der Umwelt mittels GC/ICP-MS. Unpub-

lished PhD thesis Department of Chemistry, University Essen. Feldmann J, Hirner AV (1995) Occurrence of volatile metal and metalloid species in landfill and sewage gases. Intern J

Environ Anal Chem 60, 339-359. Feldmann J, Riechmann T, Hirner AV (1996) Determination of organometallics in intra-oral air by LT-GC/ICP-MS.

Fresen J Anal Chem 354, 620-623. Feldmann J (1997) Summary of a calibration method for the determination of volatile metal(loid) compounds in envi-

ronmental gas samples by using gas chromatography - inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 12 (9), 1069-1076.

Feldmann J, Cullen WR (1997) Occurence of volatile transition metal compounds in landfill gas: synthesis of molybde-num and tungsten carbonyls in the environment. Environ Sci Technol 31 (7), 2125-2129.

Feldmann J, Kleinmann J (1997) Flüchtige Metallverbindungen im Faulgas. Korrespondenz Abwasser 44 (1), 99-104. Feldmann J, Krupp EM, Glindemann D, Hirner AV, Cullen WR (1999a) Methylated bismuth in the environment. Appl

Organomet Chem, 13 (10), 739-748. Feldmann J, Lai VWM, Cullen WR, Ma M, Lu X, Le XC (1999b) Sample preparation and storage can change arsenic

speciation in human urine. Clinical Chemistry 45 (11), 1988-1997. Feldmann J, Naels L, Haas K (2001) Cryotrapping of CO2-rich atmospheres for the analysis of volatile metal compounds

using capillary GC-ICP-MS. J Anal Atom Spectrom 16, 1040-1043. Feldmann J, Devalla S, Raab A, Hansen HR (2004) Analytical strategies for arsenic speciation in environmental and

biological samples. in Hirner AV, Emons H (eds) Organic Metal and Metalloid Species in the Environment - Analy-sis, Distribution, Processes and Toxicological Evaluation, 41-70.

160 References

Ferris MJ, Magnuson TS, Fagg JA, Thar R, Kühl M, Sheehan KB, Henson JM (2003) Microbially mediated sulphide production in a thermal, acidic algal mat community in Yellowstone National Park. Environ Microbiol 5 (10), 954-960.

Ficklin WH (1983) Separation of arsenic(III) and arsenic(V) in ground waters by ion exchange. Talanta 30, 371-373. Finn CA, Morgan LA (2002) High-resolution aeromagnetic mapping of volcanic terrain, Yellowstone National Park. J

Volcan Geotherm Res 115, 207-231. Fleck HZ (1872) Biol 8, 444. Florea AM, Dopp E, Obe G, Rettenmeier AW (2004) Genotoxicity of organometallic species. in Hirner AV, Emons H

(eds) Organic Metal and Metalloid Species in the Environment - Analysis, Distribution, Processes and Toxicological Evaluation, 205-219.

Fournier RO (1959) Death Gulch gas analysis. Yellowstone National Park, Yellowstone Research Library, vertial file. Fournier RO, Truesdell AH (1973) An empirical Na-K-Ca geothermometer for natural waters. Geochim Cosmochim

Acta 37, 1255-1275. Fournier RO, White DE, Truesdell AH (1976) Convective heat flow in Yellowstone National Park. Proceedings of the 2nd

U.N. Symposium Development and Use of Geothermal Resources, San Francisco 1975, Washington, 731-739. Fournier RO, Potter RW (1982) A revised and expanded silica (quartz) geothermometer. Geotherm Res Council Bull 11,

3-9. Fournier RO, Thompson JM, Hutchinson RA (1986) Fluctuations in composition of Cistern Spring, Norris Geyser Basin,

Yellowstone National Park, Wyoming – variable boiling and mixing 1962-1985. Extended Abstract, Fifth Interna-tional Symposium Water-Rock Interaction, Reykjavik, Iceland, 206-209.

Fournier RO (1989) Geochemistry and dynamics of the Yellowstone National Park hydrothermal system. Annu Rev Earth Pl Sc 17, 13-53.

Fournier RO, Thompson JM, Hutchinson RA (1992) The geochemistry of hot spring waters at Norris Geyser Basin, Yellowstone National Park, USA. In: Kharaka YK, Maest AS (eds) Proc. 7th Internat. Symp. Water-Rock Interac-tion, WRI-7, Park City, Utah, USA, July 13-18, 1289-1292.

Fournier RO, Thompson JM, Hutchinson RA (1994a) The geochemistry of hot spring waters at Norris Geyser Basin, Yellowstone National Park. Geothermal Resources Council Transactions 18, 177-179.

Fournier RO, Christiansen RL, Hutchinson RA, Pierce KL (1994b) A field-trip guide to Yellowstone National Park, Wyoming, Montana and Idaho - Volcanic, hydrothermal, and glacial activity in the region. US Geol Survey Bull 2099.

Fournier RO, Weltman U, Counce D, White LD, Janik CJ (2002) Results of weekly chemical and isotopic monitoring of selected springs in Norris Geyser Basin, Yellowstone National Park During June-September, 1995. US Geol Survey Open-File Report 02-344.

Fournier RO (2004) Yellowstone caldera inflation-deflation and hydrothermal Cl flux revisited. In: Wanty RB, Seal RR (eds) Water-Rock Interaction Proc of the 11th Int Symp Water-Rock Interaction, II, Balkema, 53-58.

Fowler BA (ed) (1983) Biological and environmental effects of arsenic. Topics in environmental health 6, 281. Elsevier. Francesconi KA, Edmonds JS, Hatcher BG (1988) Examination of the arsenic constituents of the herbivorous marine

gastropod Tectus pyramis: isolation of tetramethylarsonium ion. Comp Biochem Physiol 90C, 313-316. Francesconi KA, Edmonds JS (1997) Arsenic and marine organisms. Adv Inorg Chem 44, 147-189. Frankenberger WT, Karlson U (1989) Environmental factors affecting microbial production of dimethyl selenide in a

selenium-contaminated sediment. Soil Sci Soc Am J 53, 1435-1442. Freeman MC (1985) The reduction of arsenate to arsenite by an Anabaena assemblage isolated from the Waikato River.

N Z J Mar Freshwater Res 19, 277-282. Frisbee R (1960) Report on Death Gulch. Yellowstone National Park, Yellowstone Research Library, vertical file, 3. Froehlich PN, Kaul LW, Byrd JT, Andreae MO, Roe KK (1985) Arsenic, barium, germanium, tin, dimethylsulfide and

nutrient biogeochemistry in Charlotte Harbor, Florida, a phosphorus-enriched esturary. Estuar Coast Shelf S 20, 239-264.

Fujii R, Swain WC (1995) Areal distribution of selected trace elements, salinity, and major ions in shallow ground water, Tulare basin, southern San Joaquin Valley, California. US Geol Survey Water-Resour Invest Rept 95–4048.

Fuller EN, Schettler PD, Giddings JC (1966) A new method for prediction of binary gas-phase diffusion coefficients. Ind Eng Chem 58(5), 19-27.

Gailer J, George GN, Pickering IJ, Prince RC, Ringwald SC, Pemberton JE, Glass RS, Younis HS, DeYoung DW, Aposhian HV (2000) A metabolic link between arsenic and selenite: The seleno-bis(S-gluthathionyl) arsinium ion. J Am Chem Soc 122 (19), 4637-4639.

Gallagher PA, Shoemaker JA, Wei X, Brockhoff-Schwegel CA, Creed JT (2001a) Extraction and detection of arsenicals in seaweed via accelerated solvent extraction with ion chromatographic separation and ICP-MS detection. Fresen J Anal Chem 369, 71-80.

Gallagher PA, Schwegel CA, Wei X, Creed JT (2001b) Speciation and preservation of inorganic arsenic in drinking water sources using EDTA with IC separation and ICP-MS detection. J Environ Monit 3, 371-376.

Gardiner JA, Collat JW (1965) Kinetics of the stepwise hydrolysis of tetrahydroborate ion. J Am Chem Soc 87 (8), 1692-1700.

Garrott RA, Eberhardt LL, Otton JK, White PJ, Chaffee MA (2003) A geochemical trophic cascade in Yellowstone's geothermal environments. Ecosystems 5, 659-666.

References 161

Geiszinger AE, Goessler W, Francesconi KA (2002) Biotransformation of arsenate to the tetramethylarsonium ion in the marine Polychaetes Nereis diversicolor and Nereis virens. Environ Sci Technol 36, 2905-2910.

Gettar RT, Garavaglia RN, Gautier EA, Batistoni DA (2000) Determination of inorganic and organic anionic arsenic species in water by ion chromatography coupled to hydride generation–inductively coupled plasma atomic emission spectrometry. J Chromatogr A 884, 211-221.

Giggenbach WF (1975) A simple method for the collection and analysis of volcanic gas samples. B Volcanol 39, 132-145.

Giggenbach WF (1988) Geothermal solute equilibria - derivation of Na-K-Ca geoindicators. Geochim Cosmochim Acta 52, 2749-2765.

Giggenbach WF, Goguel RL (1989) Collection and analysis of geothermal and volcanic water and gas discharges. Pe-tone, New Zealand, Department of Science and Industrial Research, Chemistry Division, Report CD 3401.

Gihring TM, Druschel GK, McCleskey RB, Hamers RJ, Banfield JF (2001) Rapid arsenite oxidation by Thermus aquati-cus and Thermus thermophilus: Field and laboratory investigations. Environ Sci Technol 35 (19), 3857-3862.

Gilmour CC, Riedel GS, Ederington MC, Bell JT, Benoit JM, Gill GA, Stordal MC (1998) Methylmercury concentra-tions and production rates across a trophic gradient in the northern Everglades. Biogeochem 40, 327-345.

Gmelin L (1839) Warnung vor gewissen grünen Tapeten und Anstrichen. Karlsruher Zeit, November 24, sppl. 326. Göhler S (2003) Untersuchungen zu Wärmeentwicklung und Gasaustritten der Deutschlandschachthalde. Unpublished

master thesis, Technische Universität Bergakademie Freiberg, Department of Geology. Goff F, Janik CJ (2002) Gas geochemistry of the Valles caldera region, New Mexico and comparisons with gases at

Yellowstone, Long Valley and other geothermal systems. J Volcanol Geotherm Res 116, 299-323. Goldhahn C (2003) Behavior of arsenic in a constructed wetland facility. Unpublished master thesis, Technische Univer-

sität Bergakademie Freiberg, Department of Geology. Gómez M, Cámara C, Palacios MA, López-Gonzálvez A (1997) Anionic cartridge preconcentrators for inorganic arsenic,

monomethylarsonate and dimethylarsinate determination by on-line HPLC-HG-AAS. Fresen J Anal Chem 357, 844-849.

Gong Z, Lu X, Cullen WR, Le XC (2001) Unstable trivalent arsenic metabolites, monomethylarsonous acid and dimethy-larsinous acid. J Anal Atom Spectrom 16, 1409-1413.

Gong Z, Lu X, Ma M, Watt C, Le XC (2002) Arsenic speciation analysis. Talanta 58, 77-96. Gooch FA, Whitfield JE (1888) Analyses of waters of the Yellowstone National Park, with an account of the methods of

analysis employed. US Geol Survey Bull 47, 84. Gorby M (1994) Arsenic in Human Medicine. In: Nriagu JO (ed) Arsenic in the environment - Part II: Human Health and

Ecosystem Effects. Adv Environ Sci 26, 1-8. Gosio B (1893) Action de quelques moisissures sure les composés fixes d´arsenic. Arch Ital Biol 18, 253-265. Gosio B (1901) Recherches ultérieures sur la biologie et sur le chimisme des arsenio-moisissures. Arch Ital Biol 35, 201-

211. Gottschalk S (1997) Hydrogeologische Untersuchungen am Uran-Tailing Schneckenstein. Unpublished master thesis,

Technische Universität Bergakademie Freiberg, Department of Geology. Green HH (1918) Description of a bacterium which oxidizes arsenite to arsenate, and of one which reduces arsenate to

arsenite, isolated from a cattle-dipping tank. S Afr J Sci 14, 465-467. Greschonig H, Schmid MG, Gübitz G (1998) Capillary electrophoretic separation of inorganic and organic arsenic com-

pounds. Fresen J Anal Chem 362, 218-223. Grümping R, Mikolajczak D, Hirner AV (1998) Determination of trimethylsilanol in the environment by LT-GC/ICP-

OES and GC-MS. Fresen J Anal Chem 361, 133-139. Grüter UM, Kresimon J, Hirner AV (2000) A new HG/LT-GC/ICP-MS multi-element speciation technique for real sam-

ples in different matrices. Fresen J Anal Chem 368, 67-72. Guerin T, Molenat N, Astruc A, Pinel R (2000) Arsenic speciation in some environmental samples: a comparative study

of HG-GC-QFAAS and HPLC-ICP-MS methods. Appl Organomet Chem 14, 401-410. Gulens J, Champ DR, Jackson RE (1979) Influence of redox environments on the mobility of arsenic in ground water. In:

Jenne EA (ed) Chemical Modelling in Aqueous Systems. Am Chem Soc, 81-95. Gurzau ES, Gurzau AE (2001) Arsenic in drinking water from groundwater in Transylvania, Romania. In: Chapell WR,

Abernathy CO, Calderon RL (eds) Arsenic exposure and health effects IV. Elsevier, Amsterdam, 181–184. Haas K, Feldmann J (2000) Sampling of trace volatile metal(loid) compounds in ambient air using polymer bags: a con-

venient method. Anal Chem 72 (17), 4205-4211. Hall GEM, Pelchat JC, Gauthier G (1999) Stability of inorganic arsenic (III) and arsenic (V) in water samples. J Anal

Atom Spectrom 14, 205-214. Hambsch B, Raue B, Brauch HJ (1995) Determination of arsenic (III) for the investigation of the microbial oxidation of

arsenic (III) to arsenic (V). Acta Hydroch Hydrob 23 (4), 166-172. Hansen SH, Larsen EH, Pritzl G, Cornett C (1992) Separation of seven arsenic compounds by high-performance liquid

chromatography with on-line detection by hydrogen-argon flame atomic absorption spectrometry and inductively coupled plasma mass spectrometry. J Anal Arom Spectrom 7, 629-634.

Hansen HR, Pickford R, Thomas-Oates J, Jaspars M, Feldmann J (2004) 2-Dimethylarsinothioyl acetic acid identified in a biological sample: the first occurrence of a mammalian arsinothio(y)l metabolite. Angewandte Chemie (Int. Ed.) 43, 337-340.

162 References

Harrington JM, Fendorf SE, Rosenzweig RF (1998) Biotic generation of arsenic (III) in metal(loid) contaminated fresh-water lake sediments. Environ Sci Technol 32 (16), 2425-2430.

Hasegawa H, Sohrin Y, Matsui M, Hojo M, Kawashima M (1994) Speciation of arsenic in natural waters by solvent extraction and hydride generation atomic absorption spectrometry. Anal Chem 66, 3247-3252.

Hasegawa H, Sohrin Y, Matsui M, Takeda N, Ueda K (2002) Chemical speciation of inorganic and methylarsenic(III) compounds in aqueous solutions. Appl Organomet Chem 16, 446-450.

Healy SM, Zakharyan RA, Aposhian HV (1997) Enzymatic methylation of arsenic compounds. IV. In vitro and in vivo deficiency of the methylation of arsenite and monomethylarsonic acid in the guinea pig. Mutation Res 386, 229-239.

Heasler H, Jaworowski C (2004) Geologic overview of a bison-carcass site at Norris Geyser Basin, March 2004. Yellow-stone National Park Geology Report, 23.03.2004, 6 pages; http://www.nps.gov/yell/nature/geology/reports/ norrisbison.pdf

Heinrichs G, Udluft P (1999) Natural arsenic in Triassic rocks: a source of drinking-water contamination in Bavaria, Germany. Hydrogeol J 7, 468–476.

Henson JM, Ferris MJ (2003) Analysis of a eukaryotic microbial mat community across environmental gradients in a thermal, acidic stream. NSF Award Abstract #9977922.

Hirner AV, Feldmann J, Krupp E, Grümping R, Goguel R, Cullen WR (1998) Metal(loid)organic compounds in geo-thermal gases and waters. Org Geochem 29 (5-7), 1765-1778.

Hirner AV, Grüter UM, Kresimon J (2000) Metal(loid)organic compounds in contaminated soil. Fresen J Anal Chem 368, 263-267.

Hirner AV, Emons H (eds) (2004) Organic metal and metalloid species in the environment - analysis, distribution, proc-esses and toxicological evaluation. Springer, 328.

Hofmeister F (1894) Ueber methylirung im Thierkörper. Arch Exp Pathol Pharmakol 22, 253-260. Hopenhayn-Rich C, Smith AH, Goeden HM (1993) Human studies do not support the methylation threshold hypothesis

for the toxicity of inorganic arsenic. Environ Res 60, 161-177. Hopenhayn-Rich C, Biggs ML, Fuchs A, Bergoglio R, Tello E, Nicolli H, Smith AH (1996a) Arsenic and bladder cancer

mortality. Epidemiology 7, 557-558. Hopenhayn-Rich C, Biggs ML, Smith AH, Kalman DA, Moore LE (1996b) Methylation study of population environ-

mentally exposed to arsenic in drinking water. Environ Health Persp 104, 620-628. Horiguchi H (1970) Pollution and Poisonous Dangerous Compounds. Sankyosyuppan, Tokyo, 229. Howard AG, Arbab-Zavar MH, Apte S (1982) Seasonal variability of biological arsenic methylation in the River

Beaulieu. Mar Chem 11, 493-498. Howard AG, Arbab-Zaver MH, Apte S (1984) The behaviour of dissolved arsenic in the estuary of the River Beaulieu.

Estuar Coast Shelf S 19, 493-504. Howard AG, Hunt LE (1993) Coupled photooxidation-hydride AAS detector for the HPLC of arsenic compounds. Anal

Chem 65, 2995-2998. Howard AG, Comber SDW, Kifle D, Anatai EE, Purdie DA (1995) Arsenic speciation and seasonal changes in nutrient

availability and micro-plankton abundance in Southampton Water, U.K. Estuar Coast Shelf S 40, 435-450. Howard AG (1997) (Boro)Hydride techniques in trace element speciation. J Anal Atom Spectrom 12, 267-272. Hug SJ, Canonica L, Wegelin M, Gechter D, von Gunten U (2001) Solar oxidation and removal of arsenic at circumneu-

tral pH in iron containing waters. Environ Sci Technol 35, 2114-2121. Humphreys ED, Dueker KG, Schutt DL, Smith RB (2000) Beneath Yellowstone: evaluating plume and nonplume models

using teleseismic images of the upper mantle. GSA Today 10 (12), 1-7. Hurley JP, Krabbenhoft DP, Cleckner LB, Olson ML, Aiken GR, Rawlik PS (1998) System controls on the aqueous

distribution of mercury in the northern Florida Everglades. Biogeochemistry 40, 293-310. Intergovernmental Panel on Climate Change (1992) Climate change. In: Houghton JT, Callender BA, Varney SK (eds)

The supplementary report to the IPCC scientific assessment, Cambridge University Press, UK. Ishiguro S (1992) Industries using arsenic and arsenic compounds. Appl Organomet Chem 6, 323-331. Iyer HM, Evans JR, Zandt G, Stewart RM, Coakley JM, Roloff JN (1981) A deep low-velocity body under the Yellow-

stone caldera, Wyoming: Delineation using teleseimic P-wave residuals and tectonic interpretation. Geol Soc Am Bull, Part II, 92, 1471-1646.

Jackson CR, Langner HW, Donahoe-Christiansen J, Inskeep WP, McDermott TR (2001) Molecular analysis of microbial community structure in an arsenite-oxidizing acidic thermal spring. Environ Microbiol 3, 532-542.

Jagger TA (1899) Death Gulch - A natural bear-trap. Appelton´s popular science monthly, 54 (4), 475-481. Jamoussi B, Zafzouf M, Ben Hassine B (1996) Interferences by transition metals and their elimination by cyanide as a

complexing agent in the determination of arsenic using continuous flow HG-ICP-AES. Fresen J Anal Chem 355, 331-334.

Jantzen E (1992) Analytische Untersuchungen zur Beurteilung der Bindungsformen (Speciation) des Zinns in Elbesedi-menten unter Zuhilfenahme von GC-AAS, GC-MS, TRFA, GFAAS. Unpublished PhD thesis, University Hamburg.

Jenkins RO, Craig PJ, Goessler W, Miller D, Ostah N, Irgolic KJ (1998) Biomethylation of inorganic antimony com-pounds by an aerobic funugs: Scopulariopsis brevicaulis. Environ Sci Technol 32 (7), 882-885.

Jenkins RO, Morris TA, Craig PJ, Goessler W, Ostah N, Wills KM (2000) Evaluation of cot mattress inner foam as a potential site for microbial generation of toxic gases. Hum Exp Toxicol 19, 693-702.

References 163

Jewett KL, Brinckmann FE, Bellama JM (1975) Chemical factors influencing metal alkylation in water. in Church TM (ed) Marine Chemistry in the Coastal Environment. ACS Symp Ser 18, 304-318.

Johnson DL, Pilson MEQ (1975) The oxidation of arsenite in seawater. Environ Lett 8, 157-171. Jokai Z, Hegoczki J, Fodor P (1998) Stability and optimization of extraction of four arsenic species. Microchem J 59,

117-124. Kaise T, Ogura M, Nozaki T, Saitoh K, Sakurai T, Matsubara C, Watanabe C, Hanaoka K (1997) Biomethylation of

arsenic in an arsenic-rich freshwater environment. Appl Organomet Chem 11, 297-304. Kennedy BM, Lynch MA, Reynolds JH, Smith SP (1985) Intensive sampling of noble gases in fluids at Yellowstone: I.

Early overview of the data: Regional patterns. Geochim Cosmochim Acta 49, 1251–1261. Kennedy BM, Reynolds JH, Smith SP, Truesdell AH (1987) Helium isotopes: Lower Geyser Basin, Yellowstone Na-

tional Park. J Geophys Res 92, 12477-12489. Kharaka YK, Mariner RH, Bullen TD, Kennedy BM, Sturchio NC (1991) Geochemical investigations of hydraulic con-

nections between the Corwin Springs Known Geothermal Resources Area and adjacent parts of Yellowstone Na-tional Park. in Sorey ML (ed) Effects of potential geothermal development in the Corwin Springs Known Geother-mal Resources Area, Montana, on the thermal features of Yellowstone National Park. US Geol Survey Water-Resour Invest Rept 91-4052, Fl-F38.

Kharaka YK, Sorey ML, Thordsen JJ (2000) Large-scale hydrothermal fluid discharges in the Norris-Mammoth corridor, Yellowstone National Park, USA. J Geochem Expl 69-70, 201-205.

Kharaka YK, Thordsen JJ, White LD (2002) Isotope and chemical compositions of meteoric and thermal waters and snow from the Greater Yellowstone National Park Region. US Geol Survey Open-File Report 02-194.

Killelea DR, Aldstadt JH (2002) Identification of dimethylchloroarsine near a former herbicide factory by headspace solid-phase microextraction gas chromatography-mass spectrometry. Chemosphere 48 (9), 1003-1008.

Kim MJ, Nriagu J, Haack S (2000) Carbonate ions and arsenic dissolution by groundwater. Environ Sci Technol 34 (15), 3094-3100.

Kim MJ (2001) Separation of inorganic arsenic species in groundwater using ion exchange method. Bull Environ Contam Toxicol 67 (1), 46-51.

Kiver EP, Harris DV (1999) Geology of U.S. Parklands. 5th edition, John Wiley, New York. Kleno JG, Wolkoff P, Clausen PA, Wilkins CK, Pedersen T (2002) Degradation of the adsorbent Tenax TA by nitrogen

oxides, ozone, hydrogen peroxide, OH radical, and limonene oxidation products. Environ Sci Technol 36, 4121-4126.

Klumpp DW, Peterson PJ (1981) Chemical characteristics of arsenic in a marine food chain. Mar Biol 62 (4), 297-305. Klusman R (1991) Method and apparatus for the collection of reduced gases. United States Patent, No. 4,993,874. Knobloch T, Engewald W (1995) Sampling and gas chromatographic analysis of volatile organic compounds in hot and

extremely humid emissions. J High Resol Chromatogr 18, 635-642. Kocar BD, Garrott RA, Inskeep WP (2004) Elk exposure to arsenic in geothermal watersheds of Yellowstone National

Park, USA. Environ Toxicol Chem 23 (4), 982-989. Koch I, Feldmann J, Wang L, Andrewes P, Reimer KJ, Cullen WR (1999) Arsenic in the Meager Creek hot springs

environment, British Columbia. Sci Tot Environ 236, 101-117. Kösters J, Diaz-Bone RA, Planer-Friedrich B, Rothweiler B, Hirner AV (2003) Identification of organic arsenic, tin,

antimony and tellurium compounds in environmental samples by GC-MS. J Mol Struct, 661-662, 347-356. Korte NE, Fernando Q (1991) A review of arsenic (III) in groundwater. Crit Rev Env Contr 21 (1), 1-39. Krabbenhoft DP, Hurley JP, Olson ML, Cleckner LB (1998) Diel variability of mercury phase and species distributions

in the Florida Everglades. Biogeochemistry 40, 311-325. Krabbenhoft DP, Olson ML, Dewild JF, Clow DW, Striegl RS, Dornblaser MM, VanMetre P (2002) Mercury loading

and methymercury production and cycling in high-altitude lakes from the western United States. Water Air Soil Poll Focus 2, 233-249.

Krupp EM, Grümping R, Furchtbar URR, Hirner AV (1996) Speciation of metals and metalloids in sediments with LTGC/ICP-MS. Fresen J Anal Chem 354, 546-549.

Kumaraswamy S, Rath AK, Ramakrishnan B, Sethunathan N (2000) Wetland rice soils as sources and sinks of methane: a review and prospects for research. Biol Fertil Soils 31, 449-461.

Kuo TL (1968) Arsenic content of artesian well water in endemic area of chronic arsenic poisoning. Rep Inst Pathol National Taiwan Univ 20, 7–13.

Kutschke S (1998) Weiterführende hydrogeologische Untersuchungen an der Industriellen Absetzanlage Schneckenstein. Unpublished master thesis, Technische Universität Bergakademie Freiberg, Department of Geology.

Lai VWM, Cullen WR, Harrington CE, Reimer KJ (1997) The characterization of arsenosugars in commercially avail-able algal products including a Nostoc species of terrestrial origin. Appl Organomet Chem 11, 797-803.

Landgraf A (2002) Bestimmung des Schadstoff-Rückhaltevermögens eines natürlichen Feuchtgebietes (Schneckenstein, Vogtland). Unpublished master thesis, Technische Universität Bergakademie Freiberg, Department of Geology.

Landgraf A, Planer-Friedrich B, Merkel B (2002) Natural attenuation in a wetland under unfavorable conditions - Urani-um tailing Schneckenstein / Germany. In: Merkel B, Planer-Friedrich B, Wolkersdorfer C (eds) Uranium in the aquatic environment. Springer, 853-860.

Landner L (1989) Chemicals in the aquatic environment - Advanced Hazard Assessment. Springer, Berlin.

164 References

Langner HW, Jackson CR, McDermott TR, Inskeep WP (2001) Rapid oxidation of arsenite in a hot spring ecosystem, Yellowstone National Park. Environ Sci Technol 35, 3302-3309.

Lantzy RJ, MacKenzie FT (1979) Atmospheric trace metals: global cycles and assessment of man´s impact. Geochim Cosmochim Acta 43, 511-525.

Larsen EH, Pritzl G, Hansen SH (1993) Speciation of eight arsenic compounds in human urine by high-performance liquid chromatography with inductively coupled plasma mass spectrometric detection using antimonate for internal chromatographic standardization. J Anal Atom Spectrom 8, 557-564.

Le XC, Cullen WR, Reimer KJ (1993) Determination of urinary arsenic and impact of dietary arsenic intake. Talanta 40 (2), 185-193.

Le XC, Ma M (1997) Speciation of arsenic compounds by using ion-pair chromatography with atomic spectrometry and mass spectrometry detection. J Chromatogr A 764, 55-64.

Le XC, Ma M (1998) Short-column liquid chromatography with hydride generation atomic fluorescence detection for the speciation of arsenic. Anal Chem 70, 1926-1933.

Le XC, Lu X, Ma M, Cullen WR, Aposhian HV, Zheng B (2000a) Speciation of key arsenic metabolic intermediates in human urine. Anal Chem 72 (21), 5172-5177.

Le XC, Ma M, Lu X, Cullen WR, Aposhian HV, Zheng B (2000b) Determination of monomethylarsonous acid, a key arsenic methylation intermediate in human urine. Environ Health Persp 108 (11), 1015-1018.

Le XC, Yalcin S, Ma M (2000c) Speciation of submicrogram per liter levels of arsenic in water: on-site species separa-tion integrated with sample collection. Environ Sci Technol 34 (11), 2342-2347.

Lehmann JA, Smith RB, Schilly MM (1982) Upper crustal structure of the Yellowstone caldera from seismic delay time analyses and gravity correlations. J Geophys Res 87, 2713-2730.

Léonard A (1991) Arsenic. In: Merian E (ed) Metals and their compounds in the environment, VCH, 751-774. Lewis AJ, Palmer MR, Sturchio NC, Kemp AJ (1997) The rare earth element geochemistry of acid-sulphate and acid-

sulphate-chloride geothermal systems from Yellowstone National Park, Wyoming, USA. Geochim Cosmochim Acta 61 (4), 695-706.

Lido DR (ed) (2003) CRC Handbook of Chemistry and Physics. 84th ed. CRC Press. Lin S, Cullen WR, Thomas DJ (1999) Methylarsenicals and arsinothiols are potent inhibitors of mouse liver thioredoxin

reductase. Chem Res Toxicol 12, 924-930. Lindemann T, Prange A, Dannecker W, Neidhart B (1999) Simultaneous determination of arsenic, selenium and anti-

mony species using HPLC/ICP-MS. Fresen J Anal Chem 364, 462-466. Lindemann T, Prange A, Dannecker W, Neidhart B (2000) Stability studies of arsenic, selenium, antimony and tellurium

species in water, urine, fish and soil extracts using HPLC/ICP-MS. Fresen J Anal Chem 368, 214-220. Locke WM, Meyer GA (1994) A 12,000-year record of vertical deformation across the Yellowstone caldera margin: The

shorelines of Yellowstone Lake. J Geophys Res 99, 20079-20094. Londesborough S, Mattusch J, Wennrich R (1999) Separation of organic and inorganic arsenic species by HPLC-ICP-

MS. Fresen J Anal Chem 363, 577-581. Love JD, Good JM (1970) Hydrocarbons in thermal areas, Northwestern Wyoming. US Geol Survey Professional Paper

644-B, 23. Lowenthal DH, Pilson MEQ, Byrne RH (1977) The determination of the apparent dissociation constants of arsenic acid

in seawater. J Mar Res 35, 653-669. Luna AS, Sturgeon RE, de Campos RC (2000) Chemical vapor generation: atomic absorption by Ag, Au, Cu, and Zn

following reduction of aquo ions with sodium tetrahydroborate(III). Anal Chem 72, 3523-3531. Ma M, Le XC (1998) Effect of arsenosugar ingestion on urinary arsenic speciation. Clin Chem 44 (3), 539-550. MacKenzie FT, Lantzy RJ, Paterson V (1979) Global trace metal cycles and predictions. Math Geol 11, 99-142. Maeda S (1994) Safety and environmental effects. in Patai S (ed) The chemistry of organic arsenic, antimony, and bis-

muth compounds, John Wiley, New York, 725-759. Maillefer S, Lehr CR, Cullen WR (2003) The analysis of volatile trace compounds in landfill gases, compost heaps and

forest air. Appl Organomet Chem 17, 154-160. Mandal BK, Suzuki KT (2002) Arsenic round the world: a review. Talanta 58, 201-235. Mannigel S (2002) Hydrogeochemische Charakterisierung der Selbstreinigungsprozesse im bergbaulich beeinflußten

wetland Lengenfeld / Vogtland. Unpublished master thesis, Technische Universität Bergakademie Freiberg, Depart-ment of Geology.

Marafante E, Vather M, Norin H, Envall J, Sandström M, Christakopoulos A, Ryhage R (1987) Biotransformation of dimethylarsinic acid in mouse, hamster and man. J Appl Toxicol 7, 111-117.

Marler GD (1964) Effects of the Hebgen Lake earthquake of August 17, 1959, on the hot springs of the Firehole Geyser Basins, Yellowstone National Park. In: The Hebgen Lake, Montana earthquake of August 17, 1959. US Geol Survey Professional Paper 435, 185-198.

Marsh J (1836) Marsh arsenic test. Edinburgh New Phil J 23, 207. Mass MJ, Tennant A, Roop BC, Cullen WR, Styblo M, Thomas DJ, Kligerman AD (2001) Methylated trivalent arsenic

species are genotoxic. Chem Res Toxicol 14, 355-361. Matschullat J (2000) Arsenic in the geosphere - a review. Sci Total Environ 249, 297-312. McBride BC, Wolfe RS (1971) Biosynthesis of dimethylarsine by methanobacterium. Biochemistry 10, 4312-4317.

References 165

McCleskey RB, Nordstrom DK, Ball JW (2001) Cation-exchange separation of interfering metals from acid mine waters for accurate determination of total arsenic and arsenic(III) by hydride generation-atomic absorption spectrometry. USGS Workshop on Arsenic in the Environment, February 21-22, 2001, Denver, CO.

McCleskey RB, Nordstrom DK (2002) Water sample preservation for dissolved arsenic (III/V). 2002 Denver Annual Meeting (October 27-30, 2002), Paper No.131-6.

McCleskey RB, Nordstrom DK, Ball JW (2003) Metal interferences and their removal prior to the determination of As(T) and As(III) in acid mine waters by hydride generation atomic absorption spectrometry. US Geol Survey Wa-ter-Resour Invest Rept 03-4117.

McCleskey RB, Nordstrom DK, Maest AS (2004) Preservation of water samples for arsenic(III/V) determinations: an evaluation of the literature and new analytical results. Appl Geochem 19, 995-1009.

McSheehy S, Marcinek M, Chassaigne H, Szpunar J (2000) Identification of dimethylarsinoyl-riboside derivatives in seaweed by pneumatically assisted electrospray tandem mass spectrometry. Anal Chim Acta 410, 71-84.

Meharg AA, Hartley-Whitaker J (2002) Arsenic uptake and metabolism in arsenic resistant and nonresistant plant spe-cies. New Phytol 154, 29-43.

Merck (1983) An encyclopedia of chemicals, drugs, and biologicals. 10th ed., Rahway NJ, Merck Inc. Merkel B, Preußer R, Namoun T, Gottschalk S, Kutschke S (1998) Natural Leaching of uranium from the Schneckenstein

uranium mine tailing. In: Merkel B, Helling C (eds) Uranium Mining and Hydrogeology II. Proc Int Conf, Sven von Loga, Köln, 68-76.

Merkel B, Sperling B (1998) Hydrogeochemische Stoffsysteme, Teil II. DVWK-Schriften 117 Komissionsvertrieb Wirt-schafts- und Verlagsgesellschaft Gas und Wasser mbH, Bonn.

Mester Z, Sturgeon RE, Lam JW (2000) Sampling and determination of metal hydrides by solid phase microextraction thermal desorption inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 15 (11), 1461-1465.

Mester Z, Sturgeon R (2001) Detection of volatile arsenic chloride species during hydride generation: a new prospectus. J Anal Atom Spectrom 16 (5), 470-474.

Mester Z, Sturgeon R (2002) Detection of volatile organometal chloride species in model atmosphere above seawater and sediment. Environ Sci Technol 36 (6), 1198-1201.

Michalke K, Schramel P (1997) Coupling of capillary electrophoresis with ICP-MS for speciation investigation. Fresen J Anal Chem 357, 594-599.

Michalke K, Wickenheiser EB, Mehring M, Hirner AV, Hensel R (2000) Production of volatile derivates of metal(loid)s by microflora involved in anaerobic digestion of sewage sludge. Appl Environ Microbiol 66 (7), 2791-2796.

Moore JN, Ficklin WH, Johns C (1988) Partitioning of arsenic and metals in reducing sulfidic sediments. Environ Sci Technol 22 (4), 432-437.

Moore JN, Luoma SN (1990) Hazardous wastes from large-scale metal extraction. Environ Sci Technol 24 (9), 1279-1285.

Moore JN, Landrigan EM (1999) Mobilization of metal-contaminated sediment by ice-jam floods. Environ Geol 37 (1-2), 96-101.

Moore JN, Woessner WW (2003) Arsenic contamination in the water supply of Milltown, Montana. In: Welch AH, Stollenwerk KG (eds) Arsenic in ground water: geochemistry and cccurrence. Norwell, Massachusetts, Kluwer Aca-demic Publishers, 329-350.

Morrison GMP, Batley GE, Florence TM (1989) Metal speciation and toxicity. Chem Br 25, 791. Mortimer CE (1987) Chemie, 5.Aufl. Georg Thieme Verlag Stuttgart. Mosier AR, Delgado JA, Cochran VL, Valentine DW, Parton WJ (1997) Impact of agriculture on soil consumption of

atmospheric CH4 and a comparison of CH4 and N2O flux in subarctic, temperate and tropical grasslands. Nutr Cycl Agroecosys 49, 71-83.

Mosier AR (1998) Soil processes and global change. Biol Fertil Soils 27, 221-229. Moxon AL (1938) The effect of arsenic on the toxicity of seleniferous grains. Science 88, 81. Muffler LJP, White DE, Truesdell AH (1971) Hydrothermal explosion craters in Yellowstone National Park. Bull Geol

Soc Amer 82, 723-740. Mukai H, Ambe Y, Muku T, Takeshita K, Fukuma T (1986) Seasonal variations in methylarsenic compounds in airborne

particulate matter. Nature 324, 239-241. Mukhopadhyay R, Rosen BP, Phung LT, Silver S (2002) Microbial arsenic: from geocycles to genes and enzymes.

FEMS Microbiol Rev 26, 311-325. Murphy JB, Oppliger GL, Brimhall GH (1998) Plume-modified orogeny: An example from the western United States.

Geology 26, 731-734. Myers DJ, Heimbrook ME, Osteryoung J, Morrison SM (1973) Arsenic oxidation state in the presence of microorgan-

isms. Examination by differential pulse polarography. Environ Lett 5, 53-61. Naumov GB, Ryzhenko BN, Khodakovsky IL (1974) Handbook of thermodynamic data. Moskau. Neue H (1993) Methane emission from rice fields: Wetland rice fields may make a major contribution to global warming.

BioScience 43 (7), 466-73. New Jersey Hazardous Substances Fact Sheets (1993) Right to know program, New Jersey Department of Public Health,

Trenton, New Jersey, Denver (CO) Micromedex Inc. Newman DK, Kennedy EK, Coates JD, Ahmann D, Ellis DJ, Lovley DR, Morel FMM (1997) Dissimilatory arsenate and

sulfate reduction in Desulfotomaculum auripigmentum, sp. nov. Arch Microbiol 165, 380-388.

166 References

Niggemyer A, Spring S, Stackebrandt E, Rosenzweig RF (2001) Isolation and characterization of a novel As(V)-reducing bacterium: Implications for arsenic mobilization and the genus Desulfitobacterium. Appl Environ Mcirobiol 67 (12), 5568-5580.

Nordstrom DK (2000) Thermodynamic properties of environmental arsenic species: Limitations and needs. in Young C (ed) Minor elements 2000, processing and environmental aspects of As, Sb, Se, Te, and Bi. Society for Mining, Met-allurgy, and Exploration, 325-331.

Nordstrom DK, McCleskey RB, Ball JW (2001) Processes governing arsenic geochemistry in the thermal waters of Yellowstone National Park. USGS Workshop on Arsenic in the Environment, February 21-22, 2001, Denver, CO.

Nordstrom DK (2002) Worldwide occurrences of arsenic in ground water. Science 296 (5576), 2143-2145. Nordstrom DK (2003) Effects of microbiological and geochemical interactions in mine drainage. in Jambor JL, Blowes

DW, Ritchie AIM (eds) Environmental Aspects of Mine Wastes. Mineralogical Association of Canada 31, 227-238. Nordstrom DK, Archer DG (2003) Arsenic thermodynamic data and environmental geochemistry. in Welch AH, Stol-

lenwerk KG (eds) Arsenic in ground water: geochemistry and occurrence. Boston, Kluwer Academic Publishers, 1-26.

Nordstrom DK, McCleskey RB, Ball JW (2003) Orpiment solubility equilibrium and arsenic speciation for a hot spring at Yellowstone National Park using revised thermodynamic data. Geol Soc Am Abstracts with Programs 35 (6), 47.

Norin H, Ryhage R, Christakopoulos A, Sandström M (1983) New evidence for the presence of arsenocholine in shrimps (Pandalus borealis) by use of pyrolysis gas chromatography –– atomic absorption spectrometry/mass spectrometry. Chemosphere 12, 299-315.

Nriagu JO (ed) (1994a) Arsenic in the environment - Part I: cycling and characterization. Adv Environ Sci 26. John Wiley and Sons Inc. New York.

Nriagu JO (ed) (1994b) Arsenic in the environment - Part II: human health and ecosystem effects. Adv Environ Sci 27. John Wiley and Sons Inc. New York.

Obradovich JD (1992) Geochronology of the late Cenozoic volcanims of Yellowstone National Park and adjoining areas, Wyoming and Idaho. US Geol Survey Open-File Report 92-048.

Office of Environmental Health Hazard Assessment (1999) Acute reference exposure levels (RELs), averaging times, and toxicologic endpoints. http://www.oehha.ca.gov/air/acute_rels/pdf/7784421A.pdf.

Onishi H (1969) Arsenic. In: Wedepohl KH (ed) Handbook of geochemistry. Berlin. Springer, II(3).33. Oremland RS, Stolz JF (2003) The ecology of arsenic. Science 300, 939-944. Oscarson DW, Huang PM, Defosse C, Herbillon A (1981) Oxidative power of Mn(IV) and Fe(III) oxides with respect to

As(III) in terrestrial and aquatic environments. Nature 291, 50-51. Palacios MA, Gomez M, Camara C, Lopez MA (1997) Stability studies of arsenate, monomethylarsonate, dimethylarsi-

nate, arsenobetaine and arsenocholine in deionized water, urine and clean-up dry residue from urine samples and de-termination by liquid chromatography with microwave-assisted oxidation-hydride generation atomic absorption spectrometric detection. Anal Chim Acta 340, 209-220.

Pantsar-Kallio M, Korpela A (2000) Analysis of gaseous arsenic species and stability studies of arsine and trimethylar-sine by gas chromatography-mass spectrometry. Anal Chim Acta 410, 71-84.

Parkhurst DL, Appelo CAJ (1999) User’s guide to PHREEQC (version 2) - a computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations. US Geol Survey Water-Resour Invest Rept 99-4259.

Parris GE , Brinckman FE (1976) Reactions which relate to environmental mobility of arsenic and antimony. II. Oxida-tion of trimethylarsine and trimethylstibine. Environ Sci Technol 10, 1128-1134.

Pécheyran C, Quetel CR, Martin Lecuyer FM, Donard OFX (1998) Simultaneous determination of volatile metal (Pb,Hg, Sn, In, Ga) and nonmetal species (Se, P, As) in different atmospheres by cryofocusing and detection by ICP-MS. Anal Chem 70 (13), 2639-2645.

Pécheyran C, Lalère B, Donard OFX (2000) Volatile metal and metalloid species (Pb, Hg, Se) in a european urban at-mosphere (Bordeaux, France). Environ Sci Technol 34 (1), 27-32.

Pellizzari ED, Krost KJ (1984) Chemical transformations during ambient air sampling for organic vapors. J Anal Chem 56, 1813-1819.

Pelton JR, Smith RB (1982) Contemporary vertical surface displacements in Yellowstone National Park. J Geophys Res 87, 2745-2761.

Penrose WR (1974) Arsenic in the marine and aquatic environments: analysis, occurence and significance. CRC Crit Rev Environ Control 4, 465-482.

Pergantis SA, Winnik W, Heithmar EM, Cullen WR (1997) Investigations of arsine-generating reactions using deute-rium-labeled reagents and mass spectrometry. Talanta 44, 1941-1947.

Pergantis SA, Wangkarn S, Francesconi KA, Thomas-Oates JE (2000) Identification of arsenosugars at the picogram level using nanoelectrospray quadrupole time-of-flight mass spectrometry. Anal Chem 72, 357-366.

Petänen T, Virta M, Karp M, Romantschuk M (2001) Construction and use of broad host range mercury and arsenite sensor plasmids in the soil bacterium Pseudomonas fluorescens OS8. Microbial Ecol 40, 360-368.

Peters RL, Renesse v. Duivenbode JAD, Duyzer JH, Verhagen H (1994) The determination of terpenes in forest air. L Atmos Environ 28A, 2413-2419.

Peterson PJ, Benson LM, Zieve R (1981) Metalloids. In: Lepp NW (ed) Effect of heavy metal pollution on plants, 1: Effects of trace metals on plant function. 279-342 Applied Science Publishers.

References 167

Petito CT, Beck BD (1990) Evaluation of evidence for nonlinearities in the dose-response curve for arsenic carcinogen-isis. in Hemphull DD, Cothern CR (eds) Trace substances in environmental health XXIV. Science Reviews Limited, Northwood, 143-176.

Petrick JS, Ayala-Fierro F, Cullen WR, Carter DE, Aposhian HV (2000) Monomethylarsonous acid (MMAIII) is more toxic than arsenite in chang human hepatocytes. Toxicol Appl Pharmacol 163, 203-207.

Phillips SE, Taylor ML (1976) Oxidation of arsenite to arsenate by Alcaligenes faecalis. Appl Environ Microbiol 32, 392-39.

Pickett AW, McBride BC, Cullen WR (1988) Metabolism of trimethylarsine oxide. Appl Organomet Chem 2, 479-482. Pierce KL (1979) History and dynamics of glaciation in the northern Yellowstone National Park area. US Geol Survey

Professional Paper 729F. Pierce KL, Morgan LA (1990) The track of the Yellowstone hotspot: Volcanism, faulting, and uplift. US Geol Survey

Open-File Report 90-415. Pierce KL, Morgan LA (1992) The track of the Yellowstone hotspot: Volcanism, faulting, and uplift. in Link PK, Kuntz

MA, Platt LB (eds) Regional Geology of Eastern Idaho and Western Wyoming. Geol Soc Am Memoir 179, 1-53. Pierce KL, Cannon KP, Meyer GA (1997) Yellowtone caldera "heavy breathing" based on Yellowstone Lake and river

changes in post-glacial time. Eos 78, 802. Pierce KL, Morgan LA, Saltus RW (2000) Yellowstone plume head: postulated tectonic relations to the Vancouver slab,

continental boundaries, and climate. US Geol Survey Open-File Report 00-498. Pierce KL, Cannon KP, Meyer GA, Trebesch MJ, Watts RD (2002) Post-glacial inflation-deflation cycles, tilting, and

faulting in the Yellowstone caldera based on Yellowstone Lake shorelines. US Geol Survey Open-File Report 02-0142.

Pilson MEQ (1974) Arsenate uptake and reduction by Pocillopora verrucosa. Limnol Oceanogr 19, 339-341. Planer-Friedrich B, Matschullat J, Merkel B, Roewer G, Volke P (2002) Development of a robust technique for sampling

volatile metal(loid)s in wetlands. Anal Bioanal Chem 374 (7-8), 1191-1198. Pokrovski GS, Zakirov IV, Roux J, Testemale D, Hazemann JL, Bychkov AY, Golikova GV (2002) Experimental study

of arsenic speciation in vapor phase to 500°C: Implications for As transport and fractionation in low-density crustal fluids and volcanic gases. Geochim Cosmochim Acta 66 (19), 3453-3480.

Prohaska T, Pfeffer M, Tulipan M, Stingeder G, Mentler A, Wenzel WW (1999) Speciation of arsenic of liquid and gaseous emissions from soil in a microcosmos experiment by liquid and gas chromatograph with inductively coupled plasma mass spectrometry (ICP-MS) detection. Fresen J Anal Chem 364, 467-470.

Pyle DG, Hanan BB, Graham DW, Duncan RA (1997) Siletzia-geochemistry and geochronology of Yellowstone hot spot volcanism in a suboceanic setting. Geol Soc Am Abstracts with Programs 29 (6), A-298.

Rauret G, Rubio R, Padro A (1991) Arsenic speciation using HPLC-hydride generation (HG)-inductively coupled plasma (ICP)-atomic emission spectrometry (AES) with gas-liquid separator. Fresen J Anal Chem 340(3), 157-160.

Ravenscroft P, McArthur JM, Hoque BA (2001) Geochemical and palaeohydrological controls on pollution of groundwa-ter by arsenic. In: Chappell WR, Abernathy CO, Calderon R (eds) Arsenic Exposure and Health Effects IV, 53-78.

Reamer DC, Zoller WH (1980) Selenium biomethylation products from soil and sewage sludge. Science 208, 500-502. Reay PF, Asher CJ (1977) Preparation and purification of 74As-labeled arsenate and arsenite for use in biological experi-

ments. Anal Biochem 78, 557-560. Richardson BA (1994) Sudden infant death syndrome: a possible primary cause. J Forensic Sci Soc 34, 199-204. Ritchie JA (1961) Arsenic and antimony in New Zealand thermal waters. New Zealand Journal of Science 4, 218-229. Roberto FF, Barnes JM, Bruhn DF (2002) Evaluation of a GFP reporter gene construct for environmental arsenic detec-

tion. Talanta 58, 181-188. Robertson FN (1989) Arsenic in groundwater under oxidizing conditions, south-west United States. Environ Geochem

Health 11, 171-185. Robins RG (1985) The solubility of barium arsenate: Sherritt´s barium arsenate process. Metallurgical Transactions B,

16B, 404-406. Roscoe HE, Schorlemmer CA (1911) A treatise on chemistry. Macmillan and Co. Ltd., London, 697. Rosen BP (2002) Biochemistry of arsenic detoxification. FEBS Lett 529, 86-92. Rossum JR (1975) Checking the accuracy of water analysis through the use of conductivity. J Americ Water Works

Assoc 67, 204-205. Rowe JJ, Fournier RO, Morey GW (1973) Chemical analysis of thermal waters in Yellowstone National Park, Wyoming,

1960-65. US Geol Survey Bull 1303. Rüde TR (1996) Beiträge zur Geochemie des Arsens. Karlsruher Geochemische Hefte 19, 206. Sadiq M, Zaidi TH, Mian AA (1983) Environmental behavior of arsenic in soils: theoretical. Water Air Soil Poll 20, 369-

377. Saeki K, Sakakibara H, Sakai H, Kunito T, Tanabe S (2000) Arsenic accumulation in three species of sea turtles. Bio-

metals 13, 241-250. Sakurai T, Kaise T, Matsubara C (1998) Inorganic and methylated arsenic compounds induce cell death in murine

macrophages via different mechanisms. Chem Res Toxicol 11 (4), 273-283. Salmassi TM, Venkateswaren K, Satomi M, Nealson KH, Newman DK, Hering JG (2002) Oxidation of arsenite by Agro-

bacterium albertimagni, AOL15, sp. nov., isolated from Hot Creek, California. Geomicrobiol J 19, 53-66.

168 References

Sampayo-Reyes A, Zakharyan RA, Healy SM, Aposhian HV (2000) Monomethylarsonic acid reductase and monomethy-larsonous acid in hamster tissue. Chem Res Toxicol 13 (11), 1181-1186.

Sancha AM, Castro ML (2001) Arsenic in Latin America: occurrence, exposure, health effects and remediation. In: Chapell WR, Abernathy CO, Calderon RL (eds) Arsenic Exposure and Health Effects IV. Elsevier, Amsterdam, 87–96.

Sanders (1979) Microbial role in the demethylation and oxidation of methylated arsenicals in seawater. Chemosphere 8, 135-137.

Sanders JG (1985) Arsenic geochemistry in Chesapeake Bay: dependence upon anthropogenic inputs and phytoplankton species composition. Mar Chem 17, 329-340.

Santini JM, Sly LI, Schnagl RD, Macy JM (2000) A new chemolithoautotrophic arsenite-oxidizing bacterium isolated from a gold mine: phylogenetic, physiological, and preliminary biochemical studies. Appl Environ Microbiol 66 (1), 92-97.

Santosa SJ (2001) Arsenic and selenium species behavior during microwave-assisted conversion to their volatile hydride forming species. Anal Sci 17, i25-i28.

Sass RL (1994) Summary chapter for methane, In: Minami K, Mosier A, Sass RL (eds) CH4 and N2O global emissions and controls from rice fields and other agricultural and industrial sources. National Institute of Agro-Environmental Sciences, Yokendo, 1-7.

Savage KS, Tingle TN, O´Day PA, Wayhunas GA, Bird DK (2000) Arsenic speciation in pyrite and secondary weather-ing phases, Mother Lode Gold District, Tuolumne County, California. Appl Geochem 15, 1219-1244.

Sax NI (ed) (1986) Rapid guide to hazardous chemicals in the workplace. Van Nostrand Reinhold Co. Schall I (1995) Hydrogeologische und hydrochemische Untersuchungen im Bereich der ehemaligen Uranerzaufberei-

tungsanlage Lengenfeld (Vogtland). Unpublished master thesis Rheinisch-Westfälische Technische Hochschule Aa-chen.

Schedlbauer OF, Heumann KG (1999) Development of an isotope dilution mass spectrometric method for dimethylthal-lium speciation and first evidence of its existence in the ocean. Anal Chem 71 (24), 5459-5464.

Schlegel D, Mattusch J, Wennrich R (1996) Speciation analysis of arsenic and selenium compounds by capillary electro-phoresis. Fresen J Anal Chem 354, 535-539.

Schroeder WH (1982) Sampling and analysis of mercury and its compounds in the atmosphere. Environ Sci Technol 16, 394-400.

Schweigkofler M, Niessner R (1999) Determination of siloxanes and VOC in landfill gas and sewage gas by canister sampling and GC-MS/AES analysis. Environ Sci Technol 33 (20), 3680-3685.

Scudlark JR, Johnson DL (1982) Biological oxidation of arsenite in sea water. Estuar Coast Shelf S 14, 693-706. Seidel M (2002) Sorption von Metallen und Halbmetallen an Sedimenten im bergbaulich beeinflussten wetland Lengen-

feld / Vogtland. Freiberg Online Geosciences FOG 8; http://www.geo.tu-freiberg.de/fog/FOG_Vol_8.pdf. Seidel M, Mannigel S, Planer-Friedrich B, Merkel B (2002) Hydrogeochemical characterisation of surface water, sorp-

tion of metal(loid)s on sediments and exchange processes within the wetland Lengenfeld / Germany. In: Merkel B, Planer-Friedrich B, Wolkersdorfer C (eds) Uranium in the aquatic environment. Springer, 871-880.

Seyler P, Martin JM (1989) Biogeochemical processes affecting arsenic species distribution in a permanently stratified lake. Environ Sci Technol 23, 1258.

Shariatpanahi M, Anderson AC, Abdelghani AA, Englande AJ, Hughes J, Wilkinson RF (1981) Biotransformation of the pesticide sodium arsenate. J Environ Sci Health, Part B 16 (1), 35-47.

Shariatpanahi M, Anderson AC, Abdelghani AA, Englande AJ (1983) Microbial metabolism of an organic arsenical herbicide. in Oxley TA, Barry S (eds) Biodeterioration 5 Wiley, New York, 268-277.

Shibata Y, Morita M, Fuwa K (1992) Selenium and arsenic in biology: their chemical forms and biological functions. Adv Biophys 28, 31-80.

Shimizu N, Inoue Y, Daishima S, Yamaguchi K (1999) Liquid chromatography-mass spectrometry of arsenic compounds using the electrospray ionization with postcolumn addition of methanol. Anal Sci 15, 685-687.

Shiomi K, Kakehashi Y, Yamanaka H, Kikuchi T (1987) Identification of arsenobetaine and a tetramethylarsonium salt in the clam Meretrix lusoria. Organomet Chem 1, 177-183

Shiomi K (1994) Arsenic in marine organisms: Chemical forms and toxicological aspects. In Nriagu JO (ed) Arsenic in the environment – Part II: Human Health and Ecosystem Effects. Adv Environ Sci 27, 261-282.

Shraim A, Chiswell B, Olszowy H (1999) Speciation of arsenic by hydride generation - atomic absorption spectrometry (HG-AAS) in hydrochloric acid reaction medium. Talanta 50, 1109-1127.

Shraim A, Chiswell B, Olszowy H (2000) Use of perchloric acid as a reaction medium for speciation of arsenic by hy-dride generation-atomic absorption spectrometry. Analyst 125, 949-954.

Smedley PL, Zhang M, Zhang G, Luo Z (2001) Arsenic and other redox-sensitive elements in groundwater from the Huhhot Basin, Inner Mongolia. In: Cidu R (ed) Water-Rock Interaction 2001 Vol. 1. Swets and Zeitlinger, Lisse, 581–584.

Smedley PL, Kinniburgh DG (2002) A review of the source, behaviour and distribution of arsenic in natural waters. Appl Geochem 17, 517-568.

Smedley PL, Nicolli HB, MacDonald DMJ, Barros AJ, Tullio JO (2002) Hydrogeochemistry of arsenic and other inor-ganic constituents in groundwaters from La Pampa, Argentina. Appl Geochem 17, 259–284.

References 169

Smith AH, Lingas EO, Rahman M (2000) Contamination of drinking-water by arsenic in Bangladesh: a public health emergency. B World Health Organ 78 (9), 1093-1103.

Smith RB, Braile LW (1994) The Yellowstone hotspot. J Volcan Geotherm Res 61, 121-188. Smith RB, Siegel LJ (2000) Windows into the earth – The geological story of Yellowstone and Grand Teton National

Parks. Oxford United Press. Sohrin Y, Matsui M, Kawashima M, Hojo M, Hasegawa H (1997) Arsenic biogeochemistry affected by eutrophication in

Lake Biwa, Japan. Environ Sci Technol 31 (10), 2712-2720. Spliethoff HM, Mason RP, Hemond HF (1995) Interannual variability in the speciation and mobility of arsenic in a di-

mictic lake. Environ Sci Technol 29, 2157-2161. Spycher NF, Reed MH (1989) Evolution of a broadlands-type epithermal ore fluid along alternative p-T-paths: implica-

tions for the transport and deposition of base, precious, and volatile metals. Econ Geol 84, 328-359. Squibb KS, Fowler BA (1983) The toxicity of arsenic and its compounds. In Fowler BA (ed) Biological and environ-

mental effects of arsenic, 233-269 Amsterdam, Elsevier. Stanley WD, Hoover DB, Sorey ML, Rodriguez BD, Heran WD (1991) Electrical geophysical investigations in the Nor-

ris-Mammoth corridor, Yellowstone National Park, and the adjacent Corwin Springs known geothermal resources area. In: Sorey ML (Ed) Effects of potential geothermal development in the Corwin Springs known geothermal re-sources area, Montana, on the thermal features of Yellowstone National Park. US Geol Survey Water-Resour Invest Rept 91-4052, D1-D8.

Stauffer RE, Jenne EA, Ball JW (1980) Chemical studies of selected trace elements in hot-spring drainages of Yellow-stone National Park. US Geol Survey Professional Paper 1044-F.

Stauffer RE, Thompson JM (1984) Arsenic and antimony in geothermal waters of Yellowstone National Parl, Wyoming, USA. Geochim Cosmochim Acta 48 (2), 2547-2561.

Stein S, Levitsky A, Fateev O, Mallard G (1998) NIST mass spectral search program, Vers. 1.6d (24/08/98). Stolz JF, Ellis DJ, Blum JS, Ahmann D, Lovley DR, Oremland RS (1999) Sulfurospirillum barnesii sp. nov. and Sulfuro-

spirillum arsenophilum sp. nov., new members of the Sulfurospirillum clade of the Proteobacteria. Internat J Sys Bacteriol 49, 1177-1180.

Stummeyer J, Harazim B, Wippermann T (1996) Speciation of arsenic in water samples by high-performance liquid chromatography-hydride generation-atomic absorption spectrometry at trace levels using a post-column reaction sys-tem. Fresen J Anal Chem 354, 344-351.

Styblo M, Serves SV, Cullen WR, Thomas DJ (1997) Comparative inhibition of yeast glutathione reductase by arsenicals and arsenothiols. Chem Res Toxicol 10, 27-33.

Styblo M, Vega L, Germolec DR, Luster MI, Del Razo LM, Wang C, Cullen WR, Thomas DJ (1999a) Metabolism and toxicity of arsenicals in cultured cells. In: Chappell WR, Abernathy CO, Calderon RL (eds) Arsenic exposure and health effects, Elsevier, 311-323.

Styblo M, Del Razo L, LeCluyse EL, Hamilton GE, Wang C, Cullen WR, Thomas DJ (1999b) Metabolism of arsenic in primary cultures of human and rat hepatocytes. Chem Res Toxicol 12, 560-565.

Styblo M, Del Razo LM, Vega L, Germolec DR, LeCluyse EL, Hamilton GA, Reed W, Wang C, Cullen WR, Thomas DJ (2000) Comparative toxicity of trivalent and pentavalent inorganic and methylated arsenicals in human cells. Arch Toxicol 74, 289-299.

Sun GF, Pi JB, Li B, Guo XY, Yamauchi H, Yoshida T (2001) Progresses on researches of endemic arsenism in China: population at risk, intervention actions, and related scienti.c issues. In: Chapell WR, Abernathy CO, Calderon RL (eds) Arsenic Exposure and Health Effects IV. Elsevier, Amsterdam, 79–85.

Sur R (1999) Kopplung von gas- und flüssigkeitschromatographischen Methoden mit atom- und massenspektrometri-schen Detektoren zur Elementspeziesanalytik in biologischen Matrices am Beispiel von Arsen, Platin und Palladium. Verlag Mainz, Wissenschaftsverlag, Aachen.

Takizawa Y (1979) Epidemiology of mercury poisoning. in Nriagu JO (ed) The Biogeochemistry of Mercury in the Environment, Elsevier, 325-365.

Talmi Y, Bostick DT (1975) Determination of alkylarsenic acids in pesticide and environmental samples by gas chroma-tography with a microwave emission spectrometric detection system. Anal Chem 47 (13), 2145-2150.

Taniguchi T, Tao H, Tominaga M, Miyazaki J (1999) Sensitive determination of three arsenic species in water by ion exclusion chromatography-hydride generation-inductively coupled plasma mass spectrometry. J Anal Atom Spec-trom 14, 651-656.

Tesfalidet S, Irgum K (1988) Volatilization of arsenic as the trichloride for sample introduction in atomic spectroscopy. Anal Chem 60, 2031-2035.

Thayer JS (1995) Environmental chemistry of the heavy elements: hydrido and organo compounds. VCH Publ, Wein-heim, 145.

Thomas JE, Rhue RD (1997) Volatilization of arsenic in contaminated cattle dipping. Bull Environ Contam Toxicol 59, 882-887.

Thomas P, Sniatecki K (1995) Determination of trace amounts of arsenic species in natural waters by high-performance liquid chromatography–inductively coupled plasma mass spectrometry. J Anal Atom Spectrom 10, 615-618.

Thompson JM, Presser TS, Barnes RB, Bird DB (1975) Chemical analysis of the waters of Yellowstone National Park, Wyoming from 1965 to 1973. US Geol Survey Open-File Report 75-25.

170 References

Thompson JM, Yadav S (1979) Chemical analysis of waters from geysers, hot springs and pools in Yellowstone National Park, Wyoming from 1974 to 1978. US Geol Survey Open-File Report 79-704.

Thompson JM, Keith TEC, Consul JJ (1985) Water chemistry and mineralogy of Morgan and Growler hot springs, Las-sen KGRA, California. Transactions of the Geothermal Research Council 9, 357–362.

Thompson JM, DeMonge JM (1996) Chemical analyses of hot springs, pools, and geysers from Yellowstone National Park, Wyoming, and vicinity, 1980-1993. US Geol Survey Open-File Report 96-68.

Thornton I, Farago M (1997) The geochemistry of arsenic. in Abernathy CO, Calderon Rl, Chappell WR (eds) Arsenic Exposure and Health Effects. Chapman Hall, London, 1-16.

Tingle JB (1911) Oxidation of arsenious and antimonius oxides. J Am Chem Soc 33, 1762-1763. TrinkwV (1990) Verordnung über Trinkwasser und Wasser für Lebensmittelbetriebe (Trinkwasserverordnung vom

5.Dezember 1990). BGB1.I, Bonn, 2612-2629. Truesdell AH, Fournier RO (1976) Conditions in the deeper parts of the hot spring systems of Yellowstone National

Park, Wyoming. US Geol Survey Open-file report 76-428. Truesdell AH, Nathenson M, Rye RO (1977) The effects of subsurface boiling and dilution on the isotopic compositions

of Yellowstone thermal waters. J Geophys Res 82, 3694-3704. Truesdell AH, Thompson JM (1982) The geochemistry of Shoshone Geyser Basin, Yellowstone National Park. In

Guidebook Wyoming Geol Assoc Annu Field Conf 33rd, 153-159. Tseng WP, Chu HM, How SW, Fong JM, Lin CS, Yeh S (1968) Prevalence of skin cancer in an endemic area of chronic

arsenicism in Taiwan. J Nat Cancer Inst 40, 453–463. Tsuruta D, Hamada T, Mochida K, Nakagawa K, Kobayashi H, Ishii M (1998) Merkel cell carcinoma, Bowen's disease

and chronic occupational arsenic poisoning. Br.J. Dermatol 139, 241-244. Tucker WA, Nelken LH (1990) Diffusion coefficients in air and water. In: Lyman WJ, Reehl WF, Rosenblatt DH: Hand-

book of Chemical Property Estimation Methods. 1st ed, Am Chem Soc. Turner AW (1949) Bacterial oxidation of arsenite. Nature 164, 76-77. Ünlü MY, Fowler SW (1979) Factors affecting the flux of arsenic through the mussel Mytilus galloprovincialis. Mar Biol

51, 209-219. Vaithiyanathan P, Richardson CJ, Kavanaugh RG, Craft CB, Barkay T (1996) Relationships of euthrophication to the

distribution of mercury and to the potential for methylmercury production in the peat soils of the Everglades. Envi-ron Sci Technol 30, 2591-2597.

Van Elteren JT, Das HA, Deligny CL, Agterdenbos J, Bax D (1994) Arsenic speciation in aqueous samples using a selec-tive As(III)/As(V) preconcentration in combination with an automatable cryotrapping hydride generation procedure for monomethylarsonic acid and dimethylarsinic acid. J Radioanal Nucl Chem 179 (2), 211-219.

Van Elteren JT, Slejkovec Z (1997) Ion-exchange separation of eight arsenic compounds by high-performance liquid chromatography - UV decomposition - hydride generation - atomic fluorescence spectrometry and stability tests for food treatment procedures. J Chromatogr A 789, 339-348.

Vanifatova NG, Spivakov BY, Mattusch J, Wennrich R (1999) Separation of arsenic and selenium species by capillary zone electrophoresis in a coated capillary. J AOAC Internat 82 (6), 1587-1593.

Vather M (1999) Variation of human metabolism of arsenic. In: Chappell W.R., Abernathy C.O., Calderon R.L. (eds): Arsenic exposure and health effects 267-279.

Wagemann R (1978) Some theoretical aspects of stability and solubility of inorganic arsenic in the freshwater environ-ment. Water Res 12, 139-145.

Walsh PR, Duce RA, Fasching JL (1979) Consideration of enrichment, sources, and flux of arsenic in the troposphere. Geophys Res 84, 1719-1726.

Wang L, Huang J (1994) Chronic arsenism from drinking water in some areas of Xinjiang, China. in Nriagu JO (ed) Arsenic in the Environment, Part II: Human Health and Ecosystem Effects. John Wiley, New York, 159–172.

Wang Y, Raihala TS, Jackman AP, John RS (1996) Use of tedlar bags in VOC testing and storage: evidence of signifi-cant VOC losses. Environ Sci Technol 30 (10), 3115-3117.

Wanklyn JA (1901) Arsenic. Kegan Paul, Trench, Trübner and Co. Ltd, London, 3-4. Wassmann R, Aulakh MS (2000) The role of rice plants in regulating mechanisms of methan emissions. Biol Fertil Soils

31, 20-29. Watanabe I, Hashimoto T, Shimoyama A (1997) Methane-oxidizing activities and methanotrophic populations associated

with wetland rice plants. Biol Fertil Soils 24, 261-265. Webster JG (1999) The source of arsenic (and other elements) in the Marble-Matingao River Catchment, Mindanao-

Philippines. Geothermics 28 (1), 95-111. Webster JG, Nordstrom DK (2003) Geothermal arsenic - The source, transport and fate of arsenic in geothermal systems.

In: Welch AH, Stollwerk KG (eds) Arsenic in ground water. Kluwer Academic Publishers, Boston, 101-125. Wedepohl KH (1995) The composition of the continental crust. Geochim Cosmochim Acta 59 (7), 1217-1232. Weed WH (1889) A deadly gas-spring in the Yellowstone Park. Science 13 (315), 130-132. Weeger W, Lievremont D, Perret M, Lagarde F, Hubert JC, Leroy M, Lett MC (1999) Oxidation of arsenite to arsenate

by a bacterium isolated from an aquatic environment. Biometals 12, 141-149. Welch AH, Lico MS, Hughes JL (1988) Arsenic in ground water of the western United States. Ground Water 26 (3), 333-

347.

References 171

Welch AH, Lico MS (1998) Factors controlling As and U in shallow ground water, southern Carson Desert, Nevada. Appl Geochem 13, 521–539.

Wester RC, Maibach HI, Sedik L, Melendres J, Wade M (1993) In vivo and in vitro percutaneous absorption and skin decontamination of arsenic from water and soil. Fundam Appl Toxicol 20, 336-340.

White DE, Muffler LJP, Truesdell AH (1971) Vapor-dominated hydrothermal systems compared with hot-water systems. Econ Geol 66, 75-97.

White DE, Hutchinson RA, Keith TEC (1988) The geology and remarkable thermal activity of Norris Geyser Basin, Yellowstone National Park, Wyoming. US Geol Survey Professional Paper 1456, Washington.

Whitfield FB, Freeman DJ, Shaw KJ (1983) Trimethylarsine: an important off-flavor component in some prawn species. Chem Ind 20, 786-787.

Wickenheiser EB, Michalke K, Drescher C, Hirner AV, Hensel R (1998) Development and application of liquid and gas-chromatographic speciation techniques with element specific (ICP-MS) detection to the study of anaerobic arsenic metabolism. Fresen J Anal Chem 362, 498-501.

Wicks C, Thatcher W, Dzurisin D (1998) Migration of fluids beneath Yellowstone Caldera inferred from satellite radar interferometry. Science 282 (5388), 458-462.

Wicks CW, Thatcher W, Dzurisin D (2002) Satellite InSAR reveals a new style of deformation at Yellowstone caldera. Eos Trans AGU, 83(47), Fall Meet Suppl Abstract T12A-1291.

Wicks CW, Thatcher W, Dzurisin D (2003) Stress transfer, thermal unrest, and implications for seismic hazards associ-ated with the Norris Uplift Anomaly in Yellowstone National Park. Eos Trans AGU, 84(46), Fall Meet Suppl Ab-stract G31B-0714.

Wildfang EK, Zakharyan RA, Aposhian HV (1997) Lack of in vitro arsenite and monomethylarsonic acid methyltrans-ferase activity in the chimpanzee and squirrel monkey. The Toxicologist 36 (1,2), 34-35.

Wildfang E, Zakharyan RA, Aposhian HV (1998) Enzymatic methylation of arsenic compounds, VI.Charaterization of hamster liver arsenite and methylarsonic acid methyltransferase activities in vitro. Toxicol Appl Pharmacol 152, 366-375.

Wilhelm E, Battino R, Wilcock RJ (1977) Low-pressure solubility of gases in liquid water. Chem Rev 77, 219-262. Wilkie JA, Hering JG (1998) Rapid oxidation of geothermal arsenic(III) in streamwaters of the eastern Sierra Nevada.

Environ Sci Technol 32, 657-662. Wilkin RT, Wallschläger D, Forda RG (2003) Speciation of arsenic in sulfidic waters. Geochem Trans 4(1), 1-7. Williams M (2001) Arsenic in mine waters: an international study. Environ Geol 40 (3), 267-278. Witekowa S, Farbotko W (1972) Chemical effects of ultrasonic waves. XVII. Oxidation of arsenous acid under ultrasonic

irradiation. Acta Chim 17, 91-95. Woessner WW, Moore JN, Johns C, Popoff MA, Sartor LC, Sullivan ML (1984) Arsenic Source and Water Supply

Remedial Action Study - Milltown, Montana. Final Report, prepared for Solid Waste Bureau Montana Department of Health and Environmental Sciences, Helena, Montana.

Wongsanoon J, Boonchalermkit S, Srilachai S, Fukuda M, Nakamura S, Klusman R (1997) A novel soil gas technique applied to an arsenic contaminated area of Ron Phibun, Southern Thailand. Conference: Toxic metal studies of Pak Panang and Pattani river basins October 17. 1997, Research and Development office, Prince of Songkla University, pp. 17.

Wood JM (1974) Biological cycles for toxic elements in the environment. Science 183, 1049. Woolson EA (1983) Emissions, cycling, and effects of arsenic in soil ecosystems. In: Fowler B.A. (ed) Biological and

environmental effects of arsenic. 51-139 Elsevier. World Health Organization (1993) Guidelines for Drinking-Water Quality, 2nd ed, WHO, Geneva. Xu Y, Schoonen MAA (1995) The stability of thiosulfate in the presence of pyrite in low-temperature aqueous solutions.

Geochim Cosmochim Acta 59, 4605-4622. Xu Y, Schoonen MAA, Nordstrom DK, Cunningham KM, Ball JW (1998) Sulfur geochemistry of hydrothermal waters

in Yellowstone National Park: I. The origin of thiosulfate in hot spring waters. Geochim Cosmochim Acta 62 (23/24), 3729-3743.

Yalcin S, Le CX (2001) Speciation of arsenic using solid phase extraction cartridges. J Environ Monit 3, 81-85. Yamamoto M, Yasuda M, Yamamoto Y (1985) Hydride-generation atomic absorption spectrometry coupled with flow

injection analysis. Anal Chem 57, 1382-1385. Yamanaka K, Hayashi H, Tachikawa W, Kato K, Hasegawa A, Oku N, Okada S (1997) Metabolic methylation is a pos-

sible genotoxicity-enhancing process of inorganic arsenic. Mut Res 394, 95-101. Yamauchi H, Yamamura Y (1984a) Metabolism and excretion of orally administered dimethylarsinic acid in the ham-

sters. Toxicol Appl Pharmacol 74, 134-140. Yamauchi H, Yamamura Y (1984b) Metabolism and excretion of orally ingested trimethylarsenic in man. Bull Environ

Contam Toxicol 32(6), 682-7. Yamauchi H, Fowler BA (1994) Toxicity and metabolism of inorganic and methylated arsenicals. In Nriagu JO (ed)

Arsenic in the environment – Part II: Human Health and Ecosystem Effects. Adv Environ Sci 27, 35-43. Yang SS, Chang HL (2001) Methane emission from paddy fields in Taiwan. Biol Fertil Soils 33, 157-165. Yano Y, Miyama T, Ito Y, Yasuda T (2000) Convenient measurements and speciation of arsenic in water by use of sim-

ple pretreatments for atomic absorption spectrometry in combination with hydride generation. Anal Sci 16, 939-943.

172 References

Yin X, Hoffmann E, Lüdke C (1996) Differential determination of arsenic(III) and total arsenic with L-cysteine as prereductant using flow injection non-dispersive atomic absorption device. Fresen J Anal Chem 355, 324-326.

Yoshida K, Chen H, Inoue Y, Wanibuchi H, Fukushima S, Kuroda K, Endo G (1997) The urinary excretion of arsenic metabolites after a single oral administration of dimethylarsinic acid to rats. Arch Environ Contam Toxicol 32, 1-6.

Zakharyan RA, Aposhian HV (1999) Enzymatic reduction of arsenic compounds in mammalian systens: The rate-limiting enzyme of rabbit liver arsenic biotransformation is MMA(V) reductase. Chem Res Toxicol 12, 1278-1283.

Zakharyan RA, Ayala-Fierro F, Cullen WR, Carter DM, Aposhian HV (1999) Enzymatic methylation of arsenic com-pounds, VII. Monomethylarsonous acid (MMAIII) is ths substrate for MMA methyltransferase of rabbit liver and human hepatocytes. Toxicol Appl Pharmacol 158, 9-15.

Zhang Z, Pawliszyn J (1993) Headspace solid-phase microextraction. Anal Chem 65, 1843-1852. Zoback ML, Thompson GA (1978) Basin and Range rifting in northern Nevada: Clues from a mid-Miocene rift and its

subsequent offsets. Geology 6, 111-116.

7 APPENDIX Appendix 1 Brief description of various sites sampled for volatile metallics in a first reconnaissance study Abandoned Uranium tailings Lengenfeld und Schneckenstein / Germany Investigations started in June 2001 in wetlands at two abandoned German uranium tailings, Len-genfeld (Schall 1995) and Schneckenstein (Gottschalk 1997; Kutschke 1998; Merkel et al. 1998). The wetlands developed at the tailing dam foots and were found to have a natural attenuation effect by decreasing the metal load from tailing drainage and seepage waters. Natural attenuation is an ecologically, economically, and politically favorable alternative to continuous active waste water treatment. Wetlands are, however, also predestined for volatilization processes because of their reducing conditions. Volatilization would interfere with the positive effects of natural attenuation because of the transformation of dissolved metals in their more toxic volatile species and possible transfer through the atmosphere in the - non-monitored - vicinity of the contamination site. In Len-genfeld, aqueous As concentrations of up to several mg/L were observed to decrease rapidly to background values while passing the wetland (Mannigel 2002). Sorption to iron oxides and the residual fraction was shown to be the main attenuation process (Seidel 2002; Seidel et al. 2002). During a first survey volatile As concentrations dissolved in water were found to be between about 10-20 ng/L. Compared to values reported from literature (e.g. 0.1 to 1.8 µg/L in cattle dipping vat sites by Thomas and Rhue 1997) these values for gaseous As were low. Due to low gas sampling volumes (0.2 - 0.7 L) concentrations in the trapping solution were only 2-3 times above the detec-tion limit and, thus, uncertainties large. Further investigations with longer accumulation times could not be conducted at this site because the wetland was drained in the course of rehabilitation measures. The wetland at Schneckenstein showed only 20% removal capacity of the original dis-solved As load (Landgraf 2002; Landgraf et al. 2002). Volatile As concentrations were lower than in Lengenfeld (<1 to about 14 ng/L), probably caused by unfavorable climatic conditions with low average annual temperatures of about 5°C combined with a short growing season, low bioactivity and a reduced organic layer of only 1-3 cm thickness. US super fund site Milltown Reservoir / USA Another mine-waste influenced wetland was investigated in June 2002: The U.S. Superfund site Milltown Reservoir 10 km east of Missoula, Montana / USA. Since the construction of the reser-voir´s dam in 1907, the Clark Fork river had delivered contaminated water and sediments from Au, Cu, Pb, and Zn mining at Butte 150 km upstream and from smelter facilities at Anaconda 240 km upstream. The amount of contaminated sediment that accumulated in the Milltown Reservoir is 6.5 million tons. Four public water wells adjacent to the reservoir had to be closed in 1981 containing As concentrations of 220 to 550 µg/L (Brumbaugh et al. 1994; Moore et al. 1988; Moore and Luoma 1990; Moore and Landrigan 1999; Moore and Woessner 2003; Woessner et al. 1984). Vola-tile As concentrations sampled from the water saturated mud along the partly flooded reservoir in June 2002 were between <5 and about 250 ng/L.

174 Appendix

Geothermal waters Yellowstone National Park / USA During the same sampling campaign in June 2002 hot springs and a geothermally influenced wet-land in the Yellowstone National Park were sampled. Yellowstone National Park is located over the world´s largest continental hot spot and contains more than 10,000 geothermal features. In-creased dissolved As concentrations up to several mg/L, partly reducing, acid conditions and a large variety of microorganisms are favorable conditions for methylation and volatilization of As. Increased concentrations of methylated mercury were detected in a previous project by the U.S. Geological Survey for the geothermally influenced wetland. Volatile As concentrations sampled from the aqueous phase were between less than 10 and about 260 ng/L. Problems were encountered with the PTFE collector cells at all hot springs with temperatures >40°C because the PTFE material deformed and water breakthrough contaminated the gas trapping solution. An easier gas collection method was applied, sampling the volatile As directly from the gas phase by gas hoods over the hot spring discharges. Concentrations of volatile As in the gas phase directly above the hot springs ranged from less than 0.4 up to about 50 mg/m3. Compared to the few values on volatile As con-centrations reported so far (e.g. 50 µg/m3 for gases from waste heaps; Feldmann 1995) these values were significantly higher. Mine water irrigated rice fields / China In August 2002, three rice field sites near Luojiaqiao town / China with 1.5 and 3 month-old rice were sampled. Rice fields like natural wetlands are known to be significant CH4 emittors (Conrad 1996; Mosier et al. 1997; Mosier 1998; Neue 1993; Wassmann and Aulakh 2000; Watanabe et al. 1997; Yang and Chang 2001) making up between 20 Tg/year (Intergovernmental Panel on Climate Change 1992; Sass 1994) to 170 Tg/year (Cicerone and Oremland 1988) of a global total CH4 emission of 600 Tg/year (Kumaraswamy et al. 2000). Sediments of the Chinese rice field at Luo-jiaqiao showed increased As concentration from contaminated sediments deposited during floods from the Jangtse river as well as through irrigation with As-rich waste water from a Cu-Zn refinery. Irrigation with As-rich water is not unusual in other Asian countries neither and concentrations of soil As can reach up to 83 mg/kg (Abedin et al. 2002). The reducing conditions of the flooded rice fields were thought to be prone for release of volatile As similar to the release of CH4. Unfortu-nately sampling could be only conducted for 3 hours and no volatile As was detected in the small gas volume sampled.

Constructed wetland Jülich and coal heap Oelsnitz / Germany Investigations on a constructed reed wetland at the research center Jülich / Germany in autum 2002 (Goldhahn 2003) as well as gas sampling on a smouldering heap of As rich coal in Oelsnitz / Erzgebirge / Germany in winter 2002/2003 (Göhler 2003) yielded no detectable volatile As con-centrations, in both cases probably also owed to the inability of sampling sufficient amounts of gas.

Natural wetland Florida Everglades / USA In March 2003, 3 samples were taken from the Water Conservation Area 1 at the Loxahatchee Na-tional Wildlife Refuge in the subtropical Florida Everglades. With 600,000 ha the Everglades are one of the biggest natural wetlands. In the late 1980´s, elevated methylmercury concentrations of

Appendix 175

more than 1.5 µg/g were detected in fish and other wildlife from the Everglades. Sediments show up to 5 ng/g MeHg (Gilmour et al. 1998). Cai et al. (1997) report the occurrence of ethylmercury up to 5 ng/g in soils and sediments. Total dissolved Hg concentrations in the marshes are < 5 ng/L, MeHg up to 1 ng/L (Hurley et al. 1998). Dissolved gaseous Hg concentrations are 0.02 to 0.04 ng/L (Krabbenhoft et al. 1998). Input from increased global atmospheric Hg deposition or photo-lytic desorption and in situ methylation of Hg from peat deposits are discussed as possible Hg sources. Krabbenhoft et al. (1998) calculated an annual massflux of mercury leaving the Ever-glades of 2.2 µg/m2/year equal to only 10% of the annual input from atmospheric deposition. Eu-trophication by nutrient-enriched agricultural runoffs does not contribute significantly to total Hg built up (Gilmour et al. 1998; Vaithiyanathan et al. 1996). Similar microbial transformations as for methylated and ethylated mercury were also suspected to increase methylated and volatile As con-centrations. A previous screening by the research group of Prof. Hirner (Department for Environ-mental Analytics and Applied Geochemistry, University Essen / Germany), showed 5-630 ng/L volatile inorganic and organic As (unpublished results). The 3 samples taken in March 2003 yielded similar concentrations of about 90, 120, and 5700 ng/L total volatile As.

The database for the individual sampling sites ranges from a 1-year hydrogeochemical study in-cluding sampling of the gas phase and sequential extractions of the sediments at Lengenfeld to a 1-day sampling of only total volatile As and dissolved As speciation without the determination of further water parameters in the Chinese rice fields. In most cases, gas samples were only taken once without replicates or quality control. The results of these reconnaissance studies are thus not included in the present research work. On the following pages:

Appendix 2 Optimization of abiotic arsine generation by hydride generation from large volume solutions in a reaction vessel Note: For experimental setup see Figure 19; RV = reaction vessel, SS = stainless steel, SS-T = stainless steel Teflon (PTFE) coated, Vial = 300 mL Vial with 200 mL solution and 100 mL Headspace, HS = headspace, CC = collector cell (description of prototypes 1-3 see section 3.1), cC, VC = concentration and volume of initial solution before reaction C (control), cR, VR = concentration and volume of solution after reaction R (remaining), concentrations below detection limit (1.5 µg/L) presented as 0.3⋅DL = 0.5 µg/L; for data interpretation and calculation of % As not volatilized see section 3.2.1

Appendix 3 Optimization of trapping volatile As in liquid sorbents Note: For experimental setup see Figure 19; for conditions of arsine generation, concentrations and volumes of C-and R-samples see Appendix 2, same experiment numbers, experiments No. 52, 53, 85-105, 111, 112, 129-136, 0A-0E used solid sorbents, results shown in Appendix 5 and discussed in section 3.2.3.1; cO, VO = concentration and volume of oxi-dizing solution; SS = stainless steel, PFA = perfluoroalkoxy, PTFE = polytetrafluorethylene; T2x4 = teflon rings 2 mm inner diameter, 4 mm outer diameter, T4x6 and T6x8 analogue; for data interpretation and calculation of %As trapped see section 3.2.2

Appendix 2 equipment setup course of the experiment solutions / reagents results Ex

perim

ent N

o.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C [µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

1 31.03.01 SS 1 5 3 yes pump --- 121 --- --- --- --- 61 640 1000 950 140 1.0 30 6 2 --- --- 1.3 270 980 28 72 2 01.04.01 SS 1 5 3 yes pump --- 93 --- --- --- --- 60 320 1100 950 200 0.9 30 6 2 --- --- 1.3 530 980 50 50 3 01.04.01 SS 1 5 3 yes pump --- 87 --- --- --- --- 60 1080 1210 950 200 0.9 30 6 2 --- --- 1.5 280 980 24 76 4 02.04.01 SS 1 5 3 yes pump --- 97 --- --- --- --- 60 185 1030 950 200 0.9 30 6 2 --- --- 1.2 570 980 57 43 5 03.04.01 SS 1 5 3 yes pump --- 94 --- --- --- --- 60 915 1030 950 200 0.9 30 6 2 --- --- 1.7 290 980 29 71 6 03.04.01 SS 1 5 3 yes pump --- 151 --- --- --- --- 60 507 1060 950 200 1.0 30 6 2 --- --- 1.4 290 980 28 72 7 04.04.01 SS 1 5 3 yes pump --- 69 --- --- --- --- 60 790 1130 950 200 0.8 30 6 2 --- --- 1.6 240 980 22 78 8 04.04.01 SS 1 5 3 yes pump --- 82 --- --- --- --- 60 435 1100 950 200 0.8 30 6 2 --- --- 1.3 360 980 34 66 9 05.04.01 SS 1 5 3 yes pump --- 111 --- --- --- --- 60 864 1040 950 200 0.8 30 6 2 --- --- 1.5 230 980 23 77 10 05.04.01 SS 1 5 3 yes pump --- 38 --- --- --- --- 60 371 1090 950 100 1.1 10 15 5 --- --- 1.6 360 960 33 67 11 06.04.01 SS 1 5 3 yes pump --- 62 --- --- --- --- 60 1122 950 950 100 1.1 20 15 5 --- --- 8.4 40 970 4 96 12 06.04.01 SS 1 5 6 yes pump --- 185 --- --- --- --- 60 378 1020 950 100 1.1 20 15 5 --- --- 8.2 45 970 5 95 13 07.04.01 SS 1 5 6 yes pump --- 96 --- --- --- --- 60 1126 970 950 100 1.1 15 15 5 --- --- 2.9 260 965 27 73 14 08.04.01 SS 1 5 6 yes pump --- 144 --- --- --- --- 60 438 960 950 200 0.8 15 15 5 --- --- 1.8 110 965 12 88 15 09.04.01 SS 1 5 6 yes pump --- 286 --- --- --- --- 50 1091 970 950 100 1.2 17 15 5 --- --- 3.2 318 967 33 67 16 10.04.01 SS 1 5 6 yes pump --- 15 --- --- --- --- 60 1452 980 950 100 1.1 17 15 5 --- --- 2.8 110 967 11 89 17 11.04.01 SS 1 10 3 yes pump --- 19 --- --- --- --- 60 1452 970 950 100 1.1 17 15 5 --- --- 2.7 176 967 18 82 18 12.04.01 SS 1 10 3 yes pump --- 180 --- --- --- --- 60 1109 1030 950 50 1.4 17 15 5 --- --- 9.2 21 967 2 98 19 14.04.01 SS 1 10 3 yes pump --- 105 --- --- --- --- 60 1315 107 950 60 1.4 17 15 5 --- --- 9.0 19 967 18 82 20 15.04.01 SS 1 10 3 yes pump --- 105 --- --- --- --- 60 506 97 950 70 1.3 17 15 5 --- --- 8.8 15 967 16 84

176 A

ppendix

Appendix 2 equipment setup course of the experiment solutions / reagents results

Expe

rimen

t No.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C [µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

21 16.04.01 SS 1 5 6 yes pump --- 105 --- --- --- --- 60 1310 95 950 70 1.3 17 15 5 --- --- 8.8 11 967 12 88 22 17.04.01 SS 1 10 3 yes pump --- 105 --- --- --- --- 60 1250 99 950 70 1.3 18 15 5 --- --- 8.8 14 968 14 86 23 18.04.01 SS 1 10 3 yes pump --- 36 --- --- --- --- Var. 1190 100 950 70 1.2 20 15 5 --- --- 9.0 27 970 28 72 24 04.05.01 SS 1 10 3 yes N2 5 30 35 --- --- --- n.d. 70 77 950 100 1.2 20 15 5 --- --- --- 16 970 21 79 25 04.05.01 SS 1 10 3 no N2 10 30 20 --- --- --- n.d. 60 72 480 100 1.2 20 15 5 --- --- --- 19 500 27 73 26 05.05.01 SS 1 10 3 no N2 20 15 35 --- --- --- n.d. 70 72 480 100 1.2 20 15 5 --- --- --- 12 500 17 83 27 13.05.01 SS 2 10 6.4 no N2 10 20 60 --- --- --- 500 90 78 950 100 1.2 20 15 5 --- --- --- 5 970 7 93 28 13.05.01 SS 2 10 6.4 no N2 10 20 60 --- --- --- 800 90 59 475 100 1.2 20 15 5 --- --- --- 4 495 7 93 29 15.05.01 SS 2 10 6.4 no N2 10 20 5 20 15 20 700 90 59 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 30 16.05.01 SS 2 10 6.4 no N2 10 20 60 --- --- --- 500 90 82 475 100 1.2 20 15 5 --- --- --- 10 495 13 87 31 17.05.01 SS 2 10 6.4 no N2 10 20 0 --- 20 40 500 90 82 475 100 1.2 20 15 5 --- 20 --- 8 515 11 89 32 17.05.01 SS 2 10 6.4 no N2 10 10 0 20 10 40 500 90 82 475 100 1.2 10 15 5 20 10 --- 9 515 12 88 33 18.05.01 SS 2 10 6.4 no N2 10 20 5 20 5 20 500 80 82 475 100 1.2 20 15 5 40 6 --- 7 541 10 90 34 18.05.01 SS 2 10 6.4 no N2 5 10 0 10 10 5 200 40 77 475 50 --- 20 15 5 40 20 --- 8 555 12 88 35 20.05.01 SS 2 10 6.4 no N2 5 5 0 5 5 5 200 25 77 475 50 --- 10 15 5 20 10 --- 11 515 15 85 36 21.05.01 SS 2 10 6.4 no N2 5 10 0 10 10 10 200 45 72 475 100 1.2 20 15 5 40 20 --- 5 555 8 92 37 21.05.01 SS 2 10 6.4 no N2 5 10 0 10 10 10 200 45 72 475 100 1.2 20 15 5 40 20 --- 6 555 10 90 38 22.05.01 SS 2 10 6.4 no N2 10 20 5 20 15 20 200 90 78 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 39 23.05.01 SS 2 10 6.4 no N2 5 10 0 10 10 10 200 45 78 475 100 1.2 20 15 5 40 20 --- n.d. 555 n.d. n.d. 40 23.05.01 SS 2 10 6.4 no N2 5 10 0 0 0 20 200 35 1 450 100 1.2 10 15 5 0 0 --- 1 460 n.d. n.d.

Appendix

177


Expe

rimen

t No.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

[mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

41 11.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 73 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 42 11.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 76 475 100 1.2 20 15 5 40 20 --- 5 555 8 92 43 12.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 76 475 100 1.2 20 15 5 40 20 --- 7 555 11 89 44 12.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 73 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 45 13.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 73 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 46 13.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 74 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 47 13.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 74 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 48 15.06.01 SS-T 1 5 3 no N2 10 20 5 20 15 20 200 90 76 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 49 15.06.01 SS-T 1 10 3 no N2 10 20 5 20 15 20 200 90 76 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 50 16.06.01 SS-T 1 5 6 no N2 10 20 5 20 15 20 200 90 69 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 51 18.06.01 SS-T 1 10 6 no N2 10 20 5 20 15 20 200 90 69 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 52 18.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 75 475 100 1.2 20 15 5 40 20 --- 4 555 6 94 53 19.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 57 475 50 1.2 20 15 5 40 20 --- 2 555 4 96 54 18.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 38 475 50 1.2 20 15 5 40 20 --- 1 555 2 98 55 19.06.01 SS-T 2 10 6.4 no N2 10 20 5 20 15 20 200 90 38 475 50 1.2 20 15 5 40 20 --- 1 555 3 97

56 28.07.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 25 90 77 470 100 1.2 20 15 5 40 20 --- 13 550 20 80

57 29.07.01 SS-T 3 0.1 M no N2 10 40 15 5 0 0 var. 70 77 475 100 1.2 20 15 5 2 0 --- n.d. 497 n.d. n.d. 58 17.08.01 SS-T 3 0.1 M no N2 10 20 5 20 15 0 55 70 65 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 59 20.08.01 SS-T 3 0.1 M no N2 10 20 5 20 15 45 30 115 65 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 60 20.08.01 SS-T 3 0.1 M no N2 30 30 5 20 15 20 30 120 34 475 100 1.2 20 15 5 40 20 --- 3 555 10 90

178 A

ppendix


Expe

rimen

t No.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

[mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

61 09.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 71 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 62 09.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 71 475 100 1.2 20 15 5 40 20 --- 6 555 10 90 63 09.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 73 475 100 1.2 20 15 5 40 20 --- 4 555 6 94 64 09.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 73 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 65 10.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 50 475 50 1.2 20 15 5 40 20 --- 1 555 1 99 66 10.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 50 475 50 1.2 20 15 5 40 20 --- 1 555 1 99 67 25.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 68 27.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 25 30 95 72 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 69 27.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 455 100 1.2 20 15 5 40 20 --- 3 535 4 96 70 28.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 77 475 100 1.2 20 15 5 40 20 --- 2 555 4 96 71 28.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 77 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 72 29.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 101 488 100 1.2 20 15 5 40 20 --- 4 568 5 95 73 29.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 101 488 100 1.2 20 15 5 40 20 --- 3 568 3 97 74 30.10.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 99 475 100 1.2 20 15 5 40 20 --- 3 555 4 96 75 05.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 73 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 76 05.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 73 475 100 1.2 20 15 5 40 20 --- 3 555 5 95 77 05.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 34 475 50 1.2 20 15 5 40 20 --- 2 555 8 92 78 05.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 82 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 79 06.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 6 475 10 1.2 20 15 5 40 20 --- 2 555 36 64 80 06.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 6 475 10 1.2 20 15 5 40 20 --- 1 555 16 84

Appendix

179


Expe

rimen

t No.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C [µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

[mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

81 06.11.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 15 475 25 1.2 20 15 5 40 20 --- 1 555 5 95 82 02.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 96 475 100 1.2 20 15 5 40 20 --- 2 555 2 98

82 a 02.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 0.5 555 1 99 82 b 02.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 82 c 03.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 82 d 03.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 82 e 03.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 0.5 555 1 99 82 f 04.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 72 475 100 1.2 20 15 5 40 20 --- 0.5 555 1 99 83 04.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 96 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 84 04.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 152 475 0 1.2 20 15 5 40 20 --- 77 555 59 41 85 06.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 86 06.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 87 06.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 3 97 88 06.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 3 97 89 10.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 90 10.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 31 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 91 11.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 36 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 92 11.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 36 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 93 11.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 45 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 94 12.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 45 475 100 1.2 20 15 5 40 20 --- 4 555 12 88

180 A

ppendix

Appendix 2 equipment setup course of the experiment solutions / reagents results Ex

perim

ent N

o.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

[mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

95 13.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 44 475 100 1.2 20 15 5 40 20 --- 2 555 6 94 96 13.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 44 475 100 1.2 20 15 5 40 20 --- 1 555 3 97 97 13.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 41 475 100 1.2 20 15 5 40 20 --- 1 555 3 97 98 13.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 41 475 100 1.2 20 15 5 40 20 --- 1 555 2 98 99 14.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 43 475 100 1.2 20 15 5 40 20 --- 2 555 6 94

100 14.12.01 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 43 475 100 1.2 20 15 5 40 20 --- 2 555 5 95 101 19.02.02 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 33 475 100 1.2 20 15 5 40 20 --- 2 555 5 95 102 19.02.02 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 33 475 100 1.2 20 15 5 40 20 --- 1 555 4 96 103 24.02.02 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 66 475 100 1.2 20 15 5 40 20 --- 1 555 1 99 104 16.04.02 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 68 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 105 17.04.02 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 62 475 100 1.2 20 15 5 40 20 --- 2 555 3 97 111 04.05.03 Vial 3 0.1 M no HS --- --- --- --- --- --- --- 40 500 200 100 --- 20 15 5 --- --- --- --- 220 --- --- 112 06.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 250 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 129 18.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 130 19.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 131 19.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 132 20.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 133 20.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 134 21.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 135 21.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- --- 136 29.05.03 SS-T 3 0.1 M no N2 10 20 5 20 15 20 30 90 500 475 100 --- 20 15 5 40 20 --- --- 555 --- ---

Appendix

181


Expe

rimen

t No.

Dat

e

RV

CC

pro

toty

pe N

o.

CC

por

osity

[µm

]

CC

side

thic

knes

s [m

m]

poro

us c

eram

ic d

isk

used

gas t

rans

port

mod

e

pre-

HG

-flu

sh N

2 [m

in]

first

inje

ctio

n N

aBH

4 [m

in]

post

-NaB

H4-

flush

N2 [

min

]

seco

nd in

ject

ion

HC

l [m

in]

seco

nd in

ject

ion

NaB

H4 [

min

]

post

-HG

flus

h N

2 [m

in]

Pres

sure

[mba

r]

dura

tion

of e

xper

imen

t [m

in]

c C [µ

g/L]

VC

[mL]

1 N

HC

l [m

l] in

1 L

C so

lutio

n

pH o

f C so

lutio

n

first

inje

ctio

n N

aBH

4 [m

L] in

NaO

H

NaB

H4 %

NaO

H %

seco

nd in

ject

ion

HC

l [m

L]

seco

nd in

ject

ion

NaB

H4 [

mL]

pH a

fter r

eact

ion

c R [µ

g/L]

VR

[mL]

%A

s not

vol

atili

zed

% A

s vol

atili

zed

0A 18.02.01 SS 1 10 3 yes pump --- 27 --- --- --- --- 40 40 1000 950 100 1.1 45 3 1 --- --- 1.4 --- 995 --- --- 0B 26.02.01 SS 1 5 3 yes pump --- 100 --- --- --- --- 65 774 1000 950 100 1.3 30 6 2 --- --- --- 980 --- --- 0C 27.02.01 SS 1 5 3 yes pump --- 122 --- --- --- --- 60 623 1000 950 100 1.2 30 6 2 --- --- 1.8 --- 980 --- --- 0D 19.04.01 SS 1 10 3 yes pump --- 7 --- --- --- --- -76 1275 100 950 65 1.3 20 15 5 --- --- 9.0 --- 970 --- --- 0E 23.04.01 SS 1 10 3 yes pump --- 18 --- --- --- --- -20 1667 100 950 70 1.3 20 15 5 --- --- 9.0 --- 970 --- ---

182 A

ppendix

Appendix 3 equipment oxidation first trap oxidation second trap results

Expe

rimen

t No.

kind

of o

xida

tion

trap

oxid

izin

g su

rfac

e in

crea

sed

by

oxid

izin

g ag

ent

amou

nt o

f oxi

dizi

ng a

gent

[mL]

stre

ngth

of o

xidi

zing

age

nt

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

stre

ngth

of a

dditi

ve

pH o

xidi

zing

solu

tion

oxid

izin

g ag

ent

amou

nt o

xidi

zing

age

nt [m

L]

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

pH o

xidi

zing

solu

tion

c O [µ

g/L]

(firs

t ox

trap)

c O [µ

g/L]

(sec

ond

ox tr

ap)

VO [µ

g/L]

(firs

t ox

trap)

VO [µ

g/L]

(sec

ond

ox tr

ap)

% A

s tra

pped

(firs

t ox

trap)

% A

s tra

pped

(sec

ond

ox tr

ap)

% A

s not

trap

ped

(firs

t tra

p)

1 glass glass beads H2O2 30 10% NaOH 30 0.1 N 10.5 --- --- --- --- --- 1840 --- 60 0 16 --- 84 2 glass glass beads H2O2 30 10% NaOH 30 0.1 N 10.1 --- --- --- --- --- 490 --- 60 0 6 --- 94 3 glass glass beads H2O2 30 10% NaOH 30 0.1 N 10.3 --- --- --- --- --- 490 --- 60 0 3 --- 97 4 glass glass beads H2O2 30 10% NaOH 30 0.1 N 9.8 --- --- --- --- --- 640 --- 60 0 9 --- 91 5 glass glass beads H2O2 30 10% NaOH 30 1N 12.5 --- --- --- --- --- 530 --- 60 0 5 --- 95 6 glass glass beads H2O2 30 20% NaOH 30 1N 12.2 --- --- --- --- --- 530 --- 60 0 4 --- 96 7 glass glass beads H2O2 30 30% NaOH 30 1N 12.5 --- --- --- --- --- 2010 --- 60 0 14 --- 86 8 glass glass beads H2O2 60 30% --- --- --- 5.3 --- --- --- --- --- 1300 --- 60 0 11 --- 89 9 glass glass beads H2O2 60 10% --- --- --- 4.7 --- --- --- --- --- 1100 --- 60 0 9 --- 91 10 glass glass beads H2O2 60 20% --- --- --- 4.1 --- --- --- --- --- 1020 --- 60 0 9 --- 91 11 glass glass beads H2O2 60 20% --- --- --- 4.4 --- --- --- --- --- 1220 --- 60 0 8 --- 92 12 glass glass beads H2O2 60 20% --- --- --- 4.1 --- --- --- --- --- 620 --- 60 0 4 --- 96 13 glass glass beads H2O2 60 20% --- --- --- 4.2 --- --- --- --- --- 810 --- 60 0 7 --- 93 14 glass glass beads H2O2 60 20% --- --- --- 4.0 --- --- --- --- --- 1200 --- 60 0 9 --- 91 15 glass glass beads H2O2 30 30% NaOH 20 1N 12.6 H2O2 30 NaOH 20 11.45 690 270 50 50 6 2 94 16 glass glass beads H2O2 30 30% NaOH 20 1N 12.3 H2O2 30 NaOH 20 11.55 980 580 50 50 6 4 94 17 glass glass beads H2O2 30 30% NaOH 20 1N 12.4 H2O2 30 NaOH 20 12.23 1050 790 50 50 7 5 93 18 glass glass beads H2O2 30 30% NaOH 20 1N 12.4 H2O2 30 NaOH 20 12.32 640 370 50 50 3 2 97 19 glass glass beads H2O2 30 30% NaOH 20 1N 12.4 --- --- --- --- --- 200 --- 50 0 12 --- 88 20 glass glass beads H2O2 40 30% NaOH 10 1N 10.2 --- --- --- --- --- 260 --- 50 0 17 --- 83 21 glass glass beads H2O2 40 30% NaOH 10 1N 9.2 --- --- --- --- --- 120 --- 50 0 8 --- 92 22 glass glass beads H2O2 20 30% NaOH 30 1N 12.5 --- --- --- --- --- 150 --- 50 0 9 --- 91

Appendix

183

App.3 equipment oxidation first trap oxidation second trap results Ex

perim

ent N

o.

kind

of o

xida

tion

trap

oxid

izin

g su

rfac

e in

crea

sed

by

oxid

izin

g ag

ent

amou

nt o

f oxi

dizi

ng a

gent

[mL]

stre

ngth

of o

xidi

zing

age

nt

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

stre

ngth

of a

dditi

ve

pH o

xidi

zing

solu

tion

oxid

izin

g ag

ent

amou

nt o

xidi

zing

age

nt [m

L]

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

pH o

xidi

zing

solu

tion

c O [µ

g/L]

(firs

t ox

trap)

c O [µ

g/L]

(sec

ond

ox tr

ap)

VO [µ

g/L]

(firs

t ox

trap)

VO [µ

g/L]

(sec

ond

ox tr

ap)

% A

s tra

pped

(firs

t ox

trap)

% A

s tra

pped

(sec

ond

ox tr

ap)

% A

s not

trap

ped

(firs

t tra

p)

23 glass glass beads H2O2 30 30% NaOH 20 1N 11.9 --- --- --- --- --- 150 --- 50 0 11 --- 89 24 SS glass beads H2O2 95 30% --- --- --- --- --- --- --- --- --- 134 --- 95 0 22 --- 78 25 SS glass beads H2O2 96 30% --- --- --- --- --- --- --- --- --- 26 --- 96 0 10 --- 90 26 SS glass beads Na2S2O8 96 1.5 M --- --- --- --- --- --- --- --- --- --- --- 96 0 n.d. --- n.d. 27 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 490 --- 95 0 67 --- 33 28 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 160 --- 95 0 58 --- 42 29 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 250 --- 95 0 86 --- 14 30 SS none NaOCl 475 6-14% Cl --- --- --- --- --- --- --- --- --- 38 --- 475 0 53 --- 47 31 SS none NaOCl 190 6-14% Cl --- --- --- --- --- --- --- --- --- 100 --- 190 0 55 --- 45 32 SS none NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 272 --- 95 0 75 --- 25 33 SS glass beads NaOCl 47.5 6-14% Cl deion.H2O 95 0 --- --- --- --- --- --- 189 --- 143 0 77 --- 23 34 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 195 --- 95 0 58 --- 42 35 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 230 --- 95 0 71 --- 29 36 SS glass beads NaOCl 47.5 6-14% Cl deion.H2O 47.5 0 --- --- --- --- --- --- 232 --- 95 0 70 --- 30 37 SS glass beads NaOCl 47.5 6-14% Cl HCl 47.5 1N --- --- --- --- --- --- 2.5 --- 95 0 1 --- 99 38 SS glass beads NaOCl 95 6-14% Cl --- --- --- --- --- --- --- --- --- 305 --- 95 0 79 --- 21 39 SS glass beads NaOCl 47.5 6-14% Cl NaOH 47.5 1N --- --- --- --- --- --- 222 --- 95 0 57 --- 43 40 SS glass beads NaOCl 90 6-14% Cl --- --- --- --- --- --- --- --- --- 28 --- 90 0 n.d. --- n.d. 41 PFA none NaOCl 95 6-14 % Cl --- --- --- 11.5 --- --- --- --- --- 162 --- 95 0 45 --- 55 42 PFA T6x8 NaOCl 95 6-14 % Cl --- --- --- 11.6 --- --- --- --- --- 270 --- 95 0 77 --- 23 43 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 158 --- 95 0 47 --- 53 44 PFA T6x8, T2x4 NaOCl 140 6-14 % Cl --- --- --- --- --- --- --- --- --- 117 --- 140 0 50 --- 50

184 A

ppendix


perim

ent N

o.

kind

of o

xida

tion

trap

oxid

izin

g su

rfac

e in

crea

sed

by

oxid

izin

g ag

ent

amou

nt o

f oxi

dizi

ng a

gent

[mL]

stre

ngth

of o

xidi

zing

age

nt

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

stre

ngth

of a

dditi

ve

pH o

xidi

zing

solu

tion

oxid

izin

g ag

ent

amou

nt o

xidi

zing

age

nt [m

L]

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

pH o

xidi

zing

solu

tion

c O [µ

g/L]

(firs

t ox

trap)

c O [µ

g/L]

(sec

ond

ox tr

ap)

VO [µ

g/L]

(firs

t ox

trap)

VO [µ

g/L]

(sec

ond

ox tr

ap)

% A

s tra

pped

(firs

t ox

trap)

% A

s tra

pped

(sec

ond

ox tr

ap)

% A

s not

trap

ped

(firs

t tra

p)

45 PFA none NaOCl 260 6-14 % Cl --- --- --- --- --- --- --- --- --- 56 --- 260 0 42 --- 58 46 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 153 --- 95 0 43 --- 57 47 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 150 --- 95 0 42 --- 58 48 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 228 --- 95 0 62 --- 38 49 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 166 --- 95 0 44 --- 56 50 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 163 --- 95 0 48 --- 52 51 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 146 --- 95 0 43 --- 57 54 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 43 --- 95 0 23 --- 77 55 PFA T6x8, T2x4 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 47 --- 95 0 26 --- 74 56 PTFE T4x6 NaOCl 142.5 6-14 % Cl --- --- --- --- --- --- --- --- --- 120 --- 143 0 59 --- 41 57 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- NaOCl 95 --- --- --- 170 17 95 95 n.d. n.d. n.d. 58 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- --- --- --- --- --- 248 --- 95 0 81 --- 19 59 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- NaOCl 95 --- --- --- 163 39 95 95 53 13 47 60 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- NaOCl 95 --- --- --- 88 3 95 95 58 2 42 61 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- NaOCl 95 --- --- --- 262 31 95 95 74 9 26 62 PTFE T4x6 NaOCl 9.5 6-14 % Cl deion.H2O 85.5 0 --- NaOCl 95 --- --- --- 197 4 95 95 62 1 38 63 PTFE T4x6 NaOCl 1.9 6-14 % Cl deion.H2O 93.1 0 --- NaOCl 95 --- --- --- 293 10 95 95 86 3 14 64 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 95 --- --- --- 276 21 95 95 77 6 23 65 PTFE T4x6 NaOCl 95 6-14 % Cl deion.H2O 0 0 --- NaOCl 95 --- --- --- 162 31 95 95 66 13 34 66 PTFE T4x6 NaOCl 1.9 6-14 % Cl deion.H2O 93.1 0 --- NaOCl 95 --- --- --- 191 17 95 95 77 7 23 67 PTFE T4x6 deion.H2O 95 0 --- --- --- --- NaOCl 95 --- --- --- 3 307 95 95 1 86 99 68 PTFE T4x6 NaOCl 95 6-14 % Cl --- --- --- --- NaOCl 95 --- --- --- 302 4 95 95 89 1 11

Appendix

185


perim

ent N

o.

kind

of o

xida

tion

trap

oxid

izin

g su

rfac

e in

crea

sed

by

oxid

izin

g ag

ent

amou

nt o

f oxi

dizi

ng a

gent

[mL]

stre

ngth

of o

xidi

zing

age

nt

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

stre

ngth

of a

dditi

ve

pH o

xidi

zing

solu

tion

oxid

izin

g ag

ent

amou

nt o

xidi

zing

age

nt [m

L]

addi

tive

to o

xidi

zing

age

nt

amou

nt o

f add

itive

[mL]

pH o

xidi

zing

solu

tion

c O [µ

g/L]

(firs

t ox

trap)

c O [µ

g/L]

(sec

ond

ox tr

ap)

VO [µ

g/L]

(firs

t ox

trap)

VO [µ

g/L]

(sec

ond

ox tr

ap)

% A

s tra

pped

(firs

t ox

trap)

% A

s tra

pped

(sec

ond

ox tr

ap)

% A

s not

trap

ped

(firs

t tra

p)

69 PTFE T4x6 NaOCl 50 6-14 % Cl deion.H2O 50 0 --- NaOCl 100 --- --- --- 223 1 100 100 71 1 29 70 PTFE T4x6 NaOCl 9.5 6-14 % Cl deion.H2O 85.5 0 --- NaOCl 95 --- --- --- 290 4 95 95 78 1 22 71 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 95 --- --- --- 304 11 95 95 81 3 19 72 PTFE T4x6 NaOCl 1.9 6-14 % Cl deion.H2O 93.1 0 --- NaOCl 95 --- --- --- 455 2 95 95 92 1 8 73 PTFE T4x6 NaOCl 0.475 6-14 % Cl deion.H2O 94.53 0 --- NaOCl 95 --- --- --- 228 86 95 95 45 17 55 74 PTFE T4x6 NaOCl 0.19 6-14 % Cl deion.H2O 94.81 0 --- NaOCl 95 --- --- --- 180 62 95 95 38 13 62 75 PTFE T4x6 NaOCl 0.19 6-14 % Cl deion.H2O 94.81 0 --- NaOCl 9.5 deion.H2O 85.5 --- 166 63 95 95 48 18 52 76 PTFE T4x6 NaOCl 1.9 6-14 % Cl deion.H2O 93.1 0 --- NaOCl 9.5 deion.H2O 85.5 --- 278 5 95 95 81 1 19 77 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 0.95 deion.H2O 94.05 --- 132 1 95 95 85 1 15 78 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 0.95 deion.H2O 94.05 --- 373 1 95 95 94 1 6 79 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 0.95 deion.H2O 94.05 --- 17 1 95 95 83 2 17 80 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 0.95 deion.H2O 94.05 --- 26 1 95 95 96 2 4 81 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- NaOCl 0.95 deion.H2O 94.05 --- 72 1 95 95 100 1 0 82 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 364 --- 95 0 78 --- 22

82 a PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 283 --- 95 0 79 --- 21 82 b PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 278 --- 95 0 80 --- 20 82 c PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 285 --- 95 0 82 --- 18 82 d PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 312 --- 95 0 88 --- 12 82 e PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 296 --- 95 0 83 --- 17 82 f PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 289 --- 95 0 81 --- 19 83 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 398 --- 95 0 83 --- 17 84 PTFE T4x6 NaOCl 0.95 6-14 % Cl deion.H2O 94.05 0 --- --- --- --- --- --- 101 --- 95 0 32 --- 68

186 A

ppendix

Appendix 187

Appendix 4 Custom-made and commercial sorption tubes used for species-selective sampling of volatile As first bed second bed third bed total

sorp

tion

tube

No.

desc

riptio

n

amou

nt o

f sila

nize

d gl

ass w

ool [

mg]

kind

of s

orpt

ion

mat

eria

l

amou

nt o

f sor

ptio

n m

ater

ial [

mg]

amou

nt o

f sila

nize

d gl

ass w

ool [

mg]

kind

of s

orpt

ion

mat

eria

l

amou

nt o

f sor

ptio

n m

ater

ial [

mg]

amou

nt o

f sila

nize

d gl

ass w

ool [

mg]

kind

of s

orpt

ion

mat

eria

l

amou

nt o

f sor

ptio

n m

ater

ial [

mg]

sila

nize

d gl

ass w

ool

tota

l am

ount

sorp

tion

mat

eria

l [m

g]

tota

l am

ount

sila

nize

d gl

ass w

ool [

mg]

1 26 Tenax TA 50 13 Chromosorb 106 100 --- --- --- 15 150 54 2 26 Tenax TA 50 13 Chromosorb 105 100 --- --- --- 15 150 54 3 26 Tenax TA 50 13 Carboxen 100 --- --- --- 15 150 54 4 26 Tenax TA 50 13 Carboxen 140 --- --- --- 15 190 54 5 26 Tenax TA 50 13 Carbosieve III 150 --- --- --- 15 200 54 6 26 Tenax TA 50 13 Carbotrap 100 --- --- --- 15 150 54 7 26 Tenax TA 30 13 Chromosorb 106 120 --- --- --- 15 150 54 8 26 Tenax TA 30 13 Chromosorb 105 120 --- --- --- 15 150 54 9

double layer with Tenax

26 Tenax TA 30 13 Carbosieve S III 170 --- --- --- 15 200 54

10 26 Tenax TA 30 13 Chromosorb 106 70 13 Car-

bosieve S III

150 15 250 67

11 26 Tenax TA 30 13 Chromosorb 105 70 13 Car-

bosieve S III

150 15 250 67

12 26 Tenax TA 30 13 Carbotrap 70 13 Car-

bosieve S III

150 15 250 67

13

triple layer with Tenax

26 Tenax TA 30 13 Chromosorb 105 70 13 Car-

bosieve S III

150 15 250 67

14 26 Chromosorb 105 100 13 Carbosieve S III 150 --- --- --- 15 250 54

15 26 Chromosorb 106 100 13 Carbosieve S III 150 --- --- --- 15 250 54

16

double layer without Tenax

26 Carbotrap 100 13 Carbosieve S III 150 --- --- --- 15 250 54

17 26 Chromosorb 105 200 --- --- --- --- --- --- 15 200 41

18 26 Chromosorb 106 200 --- --- --- --- --- --- 15 200 41

19 26 Carbosieve S III 200 --- --- --- --- --- --- 15 200 41 20

single layer

26 Carbotrap 200 --- --- --- --- --- --- 15 200 41 21 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55

22 ? Tenax TA, fine mesh size ? ? unknown material,

probably Carbotrap ? --- --- --- ? ? ?

23 ? Tenax TA, fine mesh size ? ?

unknown material, probably Car-bosieve S III

? --- --- --- ? ? ?

24 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55 25 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55 26 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55 27

commercial tubes from SKC Tenax double layer

27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55 28 26 --- --- 13 --- --- --- --- --- 16 --- 55 29

double layer glass wool 26 --- --- 13 --- --- --- --- --- 16 --- 55

30 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55 31 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55

32

commercial tubes from SKC Tenax double layer 27 Tenax TA 45 13 Tenax TA 90 --- --- --- 15 135 55

188 Appendix

Appendix 5 Experiments with sorption tubes and GC-MS analysis Note: For arsine generation conditions see Appendix 2; Exp.103 tube No. 29 mounted behind tube No. 28, Exp. 104 tube 26 mounted behind 24, parallel 27 mounted behind 25; Exp. 105 tube 31 mounted behind 21, parallel 32 mounted behind 30; sorption tube numbers as explained in Appendix 4; column properties: DB-624 (L=30m, ID=0.25mm, fd=0.25µm), DB-5 (L=12m, ID=0.32mm, fd=1 µm), MDN-12 (L=30m, (from Exp. 111 on only 25 m) ID=0.25mm, fd=0.25µm))

Expe

rimen

t No.

Dat

e of

exp

erim

ent

sorp

tion

tube

No.

Dat

e of

ana

lysi

s

gas p

ress

ure

[bar

]

GC

col

umn

Des

orpt

ion

[min

/ at

°C]

Tem

pera

ture

-Tim

e-pr

ogra

m

[°C

/ m

in h

old

/ °C

/min

hea

ting

rate

]

Tota

l tim

e [m

in]

TMA

rete

ntio

n tim

e [m

in]

TMA

pea

k ar

ea

(CH

3)2A

sCl r

eten

tion

time

[min

]

(CH

3)2A

sCl p

eak

area

0A 18.02.01 18.02. 0.75 MDN12 3 / 240

40/1/4;70/2/8;140/3/15;240/1 30 7.4 6011800 11 1044700

0B 26.02.01 26.02. 0.75 DB-5 3 / 220

30/2/3;45/2/7;80/3/15;180/2 26 0.9 2470400 --- ---

0C 27.02.01 27.02. 0.75 DB-5 3 / 220

20/4/2;40/2/5;70/2/15;180/2 33 2.7 5622600 --- ---

0D 19.04.01 30 19.04. 0.75 DB-5 3 / 240

30/4/3;50/2/6;70/1/15;180/1.7 26 0.6 822510 --- ---

0E 23.04.01 30 23.04. 0.75 DB-5 3 / 240

25/4/2;40/3/3;60/2/4;80/2.8 31 0.7 286890 --- ---

52 18.06.01 26-4 18.06. 0.5 DB-5 3 / 240

20/4/2;40/3/3;60/2/4;80/2.8 34 --- --- --- ---

53 19.06.01 26-4 19.06. 0.5 DB-5 3 / 240

20/4/2;40/3/3;60/2/4;80/2.8 34 --- --- --- ---

85 06.12.01 21 12.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

86 06.12.01 9 12.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

87 06.12.01 4 11.12. 0.5 DB-624 3 / 210

20/4/2;40/3/3;60/2/4;80/2.8 34 --- --- --- ---

88 06.12.01 5 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

89 10.12.01 6 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

90 10.12.01 12 11.12. 0.5 DB-624 3 / 210

20/4/2;40/3/3;60/2/3;100/3 42 --- --- --- ---

91 11.12.01 3 11.12. 0.5 DB-624 3 / 210

20/4/2;40/3/3;60/2/3;150/4.3 60 --- --- --- ---

92 11.12.01 19 11.12. 0.5 DB-5 3 / 210

20/4/2;40/3/3;60/2/4;80/2.8 34 --- --- --- ---

93 11.12.01 20 12.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

94 12.12.01 23 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

95 13.12.01 4 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

96 13.12.01 16 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

97 13.12.01 22 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

98 13.12.01 3 13.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

99 14.12.01 12 14.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

Appendix 189

Appendix 5

Expe

rimen

t No.

Dat

e of

exp

erim

ent

sorp

tion

tube

No.

Dat

e of

ana

lysi

s

gas p

ress

ure

[bar

]

GC

col

umn

Des

orpt

ion

[min

/ at

°C]

Tem

pera

ture

-Tim

e-pr

ogra

m

[°C

/ m

in h

old

/ °C

/min

hea

ting

rate

]

Tota

l tim

e [m

in]

TMA

rete

ntio

n tim

e [m

in]

TMA

pea

k ar

ea

(CH

3)2A

sCl r

eten

tion

time

[min

]

(CH

3)2A

sCl p

eak

area

100 14.12.01 19 15.12. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

101 19.02.02 4 19.02. 0.5 DB-624 10 / 210

40/0/6;100/2/2;140/0/5;220/2 50 --- --- --- ---

102 19.02.02 12 19.02. 0.5 DB-624 3 / 210

40/2/4;70/2/3;130/1/5;180/2.5 45 --- --- --- ---

103 24.02.02 28-29 25.02. 0.5 DB-624 3 / 240

25/4/2;40/3/3;60/2/4;180/1.8 55 --- --- --- ---

104 16.04.02 24-2625-27 21.04. 0.5 DB624 3 /

210 60/0/8;100/0/2;140/0/3;200/0 45 --- --- --- ---

105 17.04.02 21-3130-32 19.04. 0.5 DB-624 3 /

240 20/4/2;40/3/3;60/2/4;100/1.8 39 --- --- --- ---

111 04.05.03 3 04.05. 0.3 MDN12 3 / 250 30/3/10;250/10 35 --- --- --- ---

06.05.03 3 06.05. 0.3 MDN12 3 / 250 40/3/10;250/10 34 --- --- --- --- 112

second desorption 18.05. 0.15 MDN12 3 / 250 40/3/10;250/10 34 3.0 4223300 6.0 3077300

18.05.03 3 18.05. 0.15 DB-5 3 / 250 40/3/10;250/10 34 3.0 3018600 --- ---

second desorption 18.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 3.2 55891000 6.0 867180

129

third desorption 18.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 3.1 4267400 --- ---

19.05.03 21 19.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 --- --- --- --- 130

second desorption 19.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 --- --- --- ---

19.05.03 3 19.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 3.0 329290000 --- --- 131

second desorption 20.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 3.0 595700 --- ---

132 20.05.03 18 20.05. 0.15 DB-5 3 / 220 40/3/10;220/10 31 2.8 492910 --- ---

133 20.05.03 17 20.05. 0.15 DB-5 3 / 220 40/3/10;220/10 31 --- --- --- ---

134 21.05.03 19 21.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 3.2 796720 --- ---

135 21.05.03 20 21.05. 0.15 DB-5 3 / 300 40/3/10;250/10 34 --- --- --- ---

136 29.05.03 Car-

boxen (new)

29.05. 0.15 DB-5 3 / 300 40/3/10;300/10 34 3.1 28189000 --- ---

190 Appendix

Appendix 6 Arsenic concentrations in oxidizing solutions mounted behind the sorption tubes during the ex-periments described in Appendix 5 to control As breakthrough Note: sorption tube No. as explained in Appendix 4; cC, VC = concentration and volume of initial solution before reaction C (control), cR, VR = concentration and volume of solution after reaction R (remaining), cO, VO = concentration and vol-ume of oxidizing solution; concentrations below detection limit (1.5 µg/L) presented as 0.3⋅DL = 0.5 µg/L;

Expe

rimen

t No.

sorp

tion

tube

No.

(k

ind

of so

rptio

n m

ater

ial)

c C [µ

g/L]

VC [m

L]

c R [µ

g/L]

VR [m

L]

c O [µ

g/L]

fir

st tu

be

c O [µ

g/L]

se

cond

tube

VO [m

L]

As %

trap

ped

behi

nd

1. so

rptio

n tu

be

As %

trap

ped

behi

nd

2. so

rptio

n tu

be

tota

l As %

trap

ped

85 21 (Tenax TA) 31 475 1 555 132 --- 95 84 ---

86 9 (Tenax TA / Carbosieve

III) 31 475 1 555 1 --- 95 0 ---

87 4 (50 mg/L Tenax TA und

140 mg/L Carboxen 31 475 1 555 1 --- 95 0 ---

88 5 (Tenax TA / Carbosieve S

III) 31 475 1 555 1 --- 95 0 --- 89 6 (Tenax TA / Carbotrap) 31 475 1 555 127 --- 95 82 ---

90 12 (Tenax TA / Carbotrap /

Carbosieve S III) 31 475 1 555 1 --- 95 0 --- 91 3 (Tenax TA / Carboxen) 36 475 1 555 1 --- 95 0 --- 92 19 (Carbosieve S III) 36 475 1 555 72 --- 95 41 --- 93 20 (Carbotrap) 45 475 1 555 124 --- 95 53 ---

94 23 (Tenax TA / Carbosieve S

III) 45 475 4 555 1 --- 95 0 --- 95 4 (Tenax TA / Carboxen) 44 475 2 555 1 --- 95 0 ---

96 16 (Carbotrap / Carbosieve S

III) 44 475 1 555 1 --- 95 0 --- 97 22 (Tenax TA / Carbotrap) 41 475 1 555 1 --- 95 0 --- 98 3 (Tenax TA / Carboxen) 41 475 1 555 1 --- 95 0 ---


Carbosieve S III) 43 475 2 555 1 --- 95 0 --- 100 19 (Carbosieve S III) 43 475 2 555 73 --- 95 36 --- 101 4 (Tenax TA / Carboxen) 33 475 2 555 5 --- 95 3 ---


Carbosieve S III) 33 475 1 555 115 --- 95 73 ---

103 28-29 (double layer glass

wool) 66 475 1 555 66 220 95 20 67 88 104 24-26 and 30-32 (Tenax TA) 68 475 2 555 110 162 95 33 49 82 105 21-31 and 30-32 (Tenax TA) 75 475 2 555 183 152 95 50 42 92

Appendix 7 Experiments with SPME (solid-phase micro extraction) technique and GC-MS analysis Note: SPME type: 1 = PDMS 100, 2 = PDMS-CAR, 3 = PDMS-CAR-DVB; column properties: DB-624 (L=30m, ID=0.25mm, fd=0.25µm), DB-5 (L=12m, ID=0.32mm, fd=1 µm), MDN-12 (L=25m, ID=0.25mm, fd=0.25µm), HP5-ms (L=30m, ID=0.25mm, fd=0.25µm; low bleed); Exp. 107-109 desorption in injection port 2 min before program start, Exp. 110 desorption in injection port 3 min before program start, all other experiments immediate program start upon injection; n.q. = not quantifyable

Appendix 7

Expe

rimen

t No.

Dat

e of

Exp

erim

ent

solu

tion

/ rea

ctio

n vi

al v

olum

e [m

L]

amou

nt o

f As [

µg]

redu

cing

age

nt

(NaB

H4 i

n N

aOH

) [m

L]

NaB

H4 %

NaO

H %

Tota

l tim

e of

exp

ositi

on [m

in]

SPM

E ty

pe

Dat

e of

Ana

lysi

s

GC

-MS

carr

ier g

as

GC

col

umn

Inje

ctio

n Te

mpe

ratu

re [°

C]

Tem

pera

ture

-Tim

e-Pr

ogra

m

[°C

/ m

in h

old

/ °C

/min

hea

ting

rate

]

Tota

l tim

e [m

in]

MM

A re

tent

ion

time

MM

A p

eak

area

DM

A re

tent

ion

time

DM

A p

eak

area

TMA

rete

ntio

n tim

e

TMA

pea

k ar

ea

(CH

3)A

sCl 2

rete

ntio

n tim

e

(CH

3)A

sCl 2

peak

are

a

(CH

3)2A

sCl r

eten

tion

time

(CH

3)2A

sCl p

eak

area

106a 15.04.03 475/ 1100 47.5 20+20 15 5 90 1 15.04.03 N2 DB-624 200 30/2/10;200/5 24 --- --- --- --- --- --- --- --- --- ---

106b 15.04.03 475/ 1100 47.5 20+20 15 5 90 2 15.04.03 N2 DB-624 200 30/2/10;200/5 24 --- --- --- --- --- --- --- --- --- ---

107 18.04.03 4/50 1 1 2 0.7 20 1 18.04.03 N2 DB-5 200 30/2/10;200/5 24 --- --- --- --- --- --- --- --- --- --- 108 19.04.03 2/50 0.5 8.5 2 0.7 20 3 19.04.03 N2 DB-5 200 30/2/10;200/5 24 --- --- --- --- --- --- --- --- --- --- 109 19.04.03 2/50 0.5 2 15 5 20 2 19.04.03 N2 DB-5 200 30/2/10;200/5 24 --- --- --- --- --- --- --- --- --- --- 110 01.05.03 12.5/26 6.25 1 2 0.7 5 2 01.05.03 N2 MDN12 250 30/3/10;250/10 30 --- --- --- --- --- --- --- --- --- --- 113 02.05.03 10/26 5 1 2 0.7 15 1 02.05.03 He HP-5ms 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- --- ---

114 02.05.03 10/26 5 1 2 0.7 15 2 02.05.03 He HP-5ms 250 40/3/10;250/10 34 --- --- 1.32 12094 1.60-2.30 2455180 --- --- --- ---

115 02.05.03 10/26 5 1 2 0.7 15 3 02.05.03 He HP-5ms 250 40/3/10;250/10 34 1.21 n.q. 1.38 272189 1.52 1.59

132618 102566 --- --- --- ---

116 05.05.03 200/300 100 1 15 5 20 1 05.05.03 He HP-5ms 250 40/3/10;250/10 34 --- --- --- --- --- --- 2.68 2.74

51759 125697 4.12 1002335

117 05.05.03 200/300 100 1 15 5 20 2 05.05.03 He HP-5ms 250 40/3/10;250/10 34 --- --- 1.41 1975 1.56 1.68

17121 212715 --- ---

2.97 3.55 3.96

78523654 15478354 6124721

118 05.05.03 200/300 100 1 15 5 20 3 05.05.03 He HP-5ms 250 40/3/10;250/10 34 1.24 788 1.41 145451 1.55 1.62

377284 254414 --- --- --- ---

Appendix

191

Appendix 7

Expe

rimen

t No.

Dat

e of

Exp

erim

ent

solu

tion

/ rea

ctio

n vi

al v

olum

e [m

L]

amou

nt o

f As [

µg]

redu

cing

age

nt

(NaB

H4 i

n N

aOH

) [m

L]

NaB

H4 %

NaO

H %

Tota

l tim

e of

exp

ositi

on [m

in]

SPM

E ty

pe

Dat

e of

Ana

lysi

s

GC

-MS

carr

ier g

as

GC

col

umn

Inje

ctio

n Te

mpe

ratu

re [°

C]

Tem

pera

ture

-Tim

e-Pr

ogra

m

[°C

/ m

in h

old

/ °C

/min

hea

ting

rate

]

Tota

l tim

e [m

in]

MM

A re

tent

ion

time

MM

A p

eak

area

DM

A re

tent

ion

time

DM

A p

eak

area

TMA

rete

ntio

n tim

e

TMA

pea

k ar

ea

(CH

3)A

sCl 2

rete

ntio

n tim

e

(CH

3)A

sCl 2

peak

are

a

(CH

3)2A

sCl r

eten

tion

time

(CH

3)2A

sCl p

eak

area

119 09.05.03 200/300 100 1 15 5 40 1 09.05.03 N2 MDN12 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- --- --- 120 09.05.03 200/300 100 1 15 5 40 2 09.05.03 N2 MDN12 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- 7.15 65464 121 13.05.03 200/300 100 1 15 5 40 2 14.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.69 n.q --- --- 5.74 683710 122 14.05.03 200/300 100 1 15 5 40 2 14.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.72 n.q --- --- 5.81 2050800 123 14.05.03 200/300 100 1 15 5 40 2 14.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.71 n.q --- --- 5.85 882660 124 14.05.03 10/26 5 1 2 0.7 40 2 14.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- 5.81 473210

2 old 17.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.61 n.q. --- --- 5.79 1962900

127a

17.05.03

200/300

100

1

7.5

2.5

40 2

(1st use)

17.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.78 851420 --- --- 5.85 5297000

2 old 20.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- 6.04 2590500

127b

17.05.03

200/300

100

1

7.5

2.5

40 2

(2nd use)

20.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- 2.7 n.q --- --- 5.84 2704600

1 17.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- --- --- 128

17.05.03

200/300

100

1

7.5

2.5

40 3 17.05.03 N2 DB-5 250 40/3/10;250/10 34 --- --- --- --- --- --- --- --- --- ---

192 A

ppendix

Appendix 193

Appendix 8 Mass spectra of different volatile arsine species from literature

AsH3

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-1 Mass spectrum for AsH3 (as published with minor variations e.g., in Stein et al. 1998 Cullen et al. 1995 Sur 1999 Pantsar-Kallio and Korpela 2000 Mester and Sturgeon 2001)

H2As-AsH2

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-2 Mass spectrum for As2H4 (as identified by Kösters et al. 2003)

150

154

75

76

78

194 Appendix

(CH3)AsH2

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-3 Mass spectrum for (CH3)AsH2 (as published with minor variations e.g., in Pergantis et al. 1997 Cullen et al. 1995 Sur 1999 Pantsar-Kallio and Korpela 2000 Mester and Sturgeon 2001)

CH3HAs-AsH2

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-4 Mass spectrum for CH3As2H3 (as identified by Kösters et al. 2003)

90 92

76

89

75

75

89

150

163 168

Appendix 195

(CH3)2AsH

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-5 Mass spectrum for (CH3)2AsH (as published with minor variations e.g., in Pergantis et al. 1997 Cullen et al. 1995 Sur 1999 Pantsar-Kallio and Korpela 2000 Mester and Sturgeon 2001)

(CH3)3As

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-6 Mass spectrum for (CH3)3AsH (as published with minor variations e.g., in Stein et al. 1998 Cullen et al. 1995 Sur 1999 Pantsar-Kallio and Korpela 2000)

120 103

75

89

90

105

89

90 106

103

75

196 Appendix

(CH3)2As-S(CH3)

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-7 Mass spectrum for (CH3)2AsS(CH3) (as identified by Kösters et al. 2003)

AsCl3

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-8 Mass spectrum for AsCl3 (as published with minor variations e.g., in Stein et al. 1998 Mester and Sturgeon 2001)

89

109

152

121

103

75

137

110 75

145

180

Appendix 197

(CH3)2AsCl

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-9 Mass spectrum for (CH3)2AsCl (as published e.g., in Mester and Sturgeon 2001)

(CH3)AsCl2

0

10

20

30

40

50

60

70

80

90

100

70 80 90 100 110 120 130 140 150 160 170 180m/z

rel.

inte

nsity

[%]

Appendix 8-10 Mass spectrum for (CH3)AsCl2 (as published e.g., in Mester and Sturgeon 2001)

105

75

140 89

125

110

75

145

160

89

125

198 Appendix

Appendix 9 Explanation to the digital database for Yellowstone sampling 2003

Site information, pictures, and all data gathered from the sampling of Yellowstone geothermal ar-eas in 2003 are presented in an Access database (Yellowstone2003.mdb) on the enclosed CD. Opening the Access file, you see a welcome screen that you can confirm with OK to get to the main menu:

Here you have 3 options:

• Level 1 (if you are not familiar with Access): click on the Yellowstone Arch picture to enter the Main_Data_Form and get a site-specific overview over the gathered data that does not require any pre-knowledge in Access. This main data form will be explained in detail in the following and contains a combination of all available data).

• Level 2: If you are familiar with Access, and are looking for a special information you can di-rectly proceed to a table, query, or form object by double click on the scroll list below

• Level 3 (advanced users only): open the usual MS Access user interface with the display of all tables, queries, and forms, where you can edit or create new tables, queries, or forms.

The Main Data Form opens with 6 pages: “general information”, “water”, “Gas - total conc.”, “Gas comparison wGH-nGH”, “Gas - total conc. (P, DB)”, and “Gas - As speciation”. The record bar on the bottom shows 48 records in alphabetical order for the 48 sites sampled (8 sites at Geyser Spring Group, Gibbon Geyser Basin; 3 sites at Hazle Lake; 3 sites in the Pocket Basin, Lower Gey-ser Basin; 6 sites in the Nymph Lake / Frying Pan Spring area; 28 sites at Ragged Hills, Norris Geyser Basin). The information listed for each record is linked to different columns of different tables under Objects / Tables in the Access Menu. These source information are explained in the following. Throughout the database mouseover provides additional information.

Appendix 199

First page: General Information: The primary key (the unique link, under which all information can be defined unambiguously) for all information on this page is the Site_Code. The bold area name on top stems from the column “Area” in the table “Names_Descriptions” ([table]!column = [Names_Descriptions]!Area).

Site Code [Names_Descriptions]!Site_Code; Code abbreviation: YNP = Yellowstone Na-tional Park; Area (GG = Gibbon Geyser Basin, HL = Hazle Lake, LG = Lower Geyser Basin, NL = Nymph Lake, RH = Ragged Hills); No. Sampling Site

Site Name [Names_Descriptions]!Site_Name; names in quotation marks are unofficial Longitude [Coordinates_Altitudes_calc]!Longitude; value is a calculated value, according

to the following query [Query: calc- Coordinates_Altitudes]: If([Coordinates-Altitudes-DiffGPS]!Site_Code=[Coordinates-Altitudes]!Site_Code;[Coordinates-

Altitudes-DiffGPS]!Longitude_calcTNTmips;if([Coordinates-Altitudes]!Longitude2nd=-

999;[Coordinates-Altitudes]!Longitude1st;if([Coordinates-Altitudes]!Longitude1st=-

999;[Coordinates-Altitudes]!Longitude2nd;([Coordinates-Altitudes]!Longitude1st+[Coordinates-

Altitudes]!Longitude2nd)/2)))

In the Ragged Hills area several sites were surveyed by Differential GPS. Since this information is more accurate than normal GPS measurements, it was taken whenever available. The horizontal precision of the differential GPS is given in [Coordinates-Altitudes-DiffGPS]!horizontal_precision_m. [Coordinates-Altitudes-DiffGPS]!Easting contains the original UTM values (Zone 12 North, spheroid GRS80, datum NAD83). The longitude (datum WGS 72) was calcu-lated from the UTM coordinates with the GIS program TNTmips 6.8 from Mi-

200 Appendix

croImages ([Coordinates-Altitudes-DiffGPS]!Longitude_calcTNTmips). When no information from differential GPS was available, data obtained from a GPS GARMIN-12 were taken ([Coordinates-Altitudes]!Longitude1st and [Coordi-nates-Altitudes]!Longitude2nd). In the case of two measurements the mean value was calculated.

Latitude [Coordinates_Altitudes_calc]!Latitude (calculated analogous to the longitude) Altitude [Coordinates_Altitudes_calc]!Altitude_m; similar calculation like for longitude

and latitude; vertical precision indicated in [Coordinates-Altitudes-DiffGPS]!vertical_precision_m; mean altitude in feet is given in [Coordi-nates_Altitudes_calc]!Altitude_feet (feet ⋅ 0.3048 = meter)

Site Description [Names_Descriptions]!Site_Description Left picture [Pictures]!Picture1 contains the file name of each picture that is linked via a Vis-

ual Basic Script to and will be displayed together with each record; physically the pictures are saved in the folder Figures / Database-Yellowstone2003; double-click on the thumbnail picture opens an enlarged version of the same picture ([Pictures]! Picture1G)

Right picture [Pictures]!Picture2 (thumbnail); [Pictures]!Picture2G (enlarged version) Available information for this site (with sampling date 2003)

The fields herein are clustered in 5 sections corresponding to the 5 pages on top. The list gives an overview where information can be expected for each sampling site. The fields contain sampling dates, whenever samples were taken, otherwise the field (and also the corresponding page) is blank. Mouseover gives further in-formation, e.g., on what “standard parameters” are (field information from table [Summary-available-site-information]!)

Second page: Water Data on this page are linked to the subform “Water-subform”. This subform contains 62 records for 41 sites, 7 of the sampling sites had mainly fumarole activity and could not be sampled for water. Some of the sites were sampled not only once, but 2 or 3 times. primary key Analysis-code-water ([Analysis-code-water]!Analysis-code-water); the analysis

code water consists of the site code, the No. of the sampling campaign (1 = June, 5th, to June, 21st, 2003; 2 = June, 30th, to July, 12th, 2003; 3 = September, 3rd, to September, 15th, 2003; 4 = September 21st, to October, 9th, 2003), and W for wa-ter (e.g., YNPGG04-1W is the first water sampling campaign for site YNPGG04, YNPGG04-2W the second, etc.)

Note: The record bar within the subform “water” shows 1 of 3 records, indicating that sampling site YNPRH04 was sampled 3 times. Arrows to the left and right can be used to scroll within the subform water. To go to the next sampling site, the lower record bar (of the Main Data Form) has to be used. The page “water” is blank when there are no water data. Missing values are administered as -999

Appendix 201

(e.g., Mercury - Hg); values below detection limit as negative values (e.g., <0.06 N(V) = -0.06). Mouseover on values shows sample preparation, analytical meth-ods and devices.

number of sampling campaign [Analysis-code-water]!Sampling_No Sampling date [Analysis-code-water]!Sampling_Date Sampling time [Analysis-code-water]!Sampling_Time

On-site physicochemical parameters: pH [On-site-pH]!pH_corrected (= field pH; field pH was not accepted when the ana-

lytical error exceeded 5%, then either the laboratory pH was taken or pH was adapted to best fit from speciation modeling within a range of ± 0.1 pH units)

conductivity [On-site-conductivity]!conductivity_uS/cm Temperature [On-site-Temperatures-calc]!Temperature_Mean_°C; the value was calculated as

the mean of temperature read from pH ([On-site-Temperature]!Temperature_pH_°C) and conductivity meter ([On-site-Temperature]!Temperature_conductivity_°C), except when one of the values was not accepted, e.g., due to device failure [Query: calc-Temperatures]: If([On-site-Temperatures-calc]!Accept_temp_cond=false;[On-site-

Temperature]!Temperature_pH_°C;If([On-site-Temperatures-calc]!Accept_temp_pH=False;[On-

site-Temperature]!Temperature_conductivity_°C;([On-site-

Temperature]!Temperature_conductivity_°C+[On-site-Temperature]!Temperature_pH_°C)/2))

Redox potential [On-site-Redoxpotential-calc]!EH_corrected_mV; redox potential was measured in the field against the Ag/AgCl-electrode; [On-site-Redoxpotential-

202 Appendix

calc]!EH_AgAgCl-Potential_mV) and corrected for standard hydrogen potential and temperature [Query: calc-Redoxpotential]: If([On-site-Redoxpotential]![EH_AgAgCl-Potential_mV]=-999;-999;([On-site-

Redoxpotential]![EH_AgAgCl-Potential_mV]+(-0.7443⋅[On-site-Temperatures-

calc]!Temperature_Mean_°C+224.98)))

pe value was calculated from the corrected EH value ([On-site-Redoxpotential-calc]!pe) according to the following query [Query: calc-Redoxpotential]: If([On-site-Redoxpotential]![EH_AgAgCl-Potential_mV]=-999;-999;16.9⋅[On-site-

Redoxpotential-calc]!EH_corrected_mV/1000) Oxygen [On-site-Oxygen]!O2_%, and [On-site-Oxygen]!O2_mg/L Due to problems with the device (especially at temperatures > 50°C), Oxygen

was measured only in 22 samples; these data had to be attached via a sub-subform [Form On-site-Oxygen]; in the water subform data fields will be empty when no data exist

On-site photometry: SiO2 [On-site-photometry-SiO2]!SiO2_mgL, SiO2 QC photometry / ICP-AES, default: invisible, click on the option displays

mean value and range of analysis (ROA) of values from on-site photometry compared to those from ICP-AES for a quality check (table [Comparison-Si-determination]! based on the Query calc-Comparison-Si)

S(-II) [On-site-photometry-S2-]!S2-_mgL Fe [On-site-photometry-Fe]!Fe_tot_mgL and [On-site-photometry-Fe]!Fe_II_mgL: [On-site-photometry-Fe-calc]!Fe_III_mgL, [On-site-photometry-Fe-

calc]!Fe_II_%, and [On-site-photometry-Fe-calc]!Fe_III_%: values calculated as difference and percentage [Query: calc-Photometry-Fe]

Fe QC photometry / ICP-AES, default: invisible, click on the option displays mean value and range of analysis (ROA) of values from on-site photometry compared to those from ICP-AES for a quality check (table [Comparison-Fe-determination]! based on the Query: calc-Comparison-Fe)

Arsenic speciation: Total As (AAS) [Comparison-As-determination]!Total-As-AAS (essentially the same as [AAS-

Freiberg-water-As_total]!As_final_concentration_ugL, but rounded down to 2 decimals); raw data in the same table: extinction_1st_determination, extinc-tion_2nd_determination, extinction_mean; concentrations derived via Excel cali-bration curves (not shown in the Access database) are in the column As_concentration_ugL; column As_final_concentration_ugL considers dilution ratios

As(III) [AAS-Freiberg-water-As-species]!AsIII_ugL and [AAS-Freiberg-water-As-species]!AsIII_% [Query: calc-Arsenic-species]

Appendix 203

(raw data as explained for total As in table AAS-Freiberg-water-AsIII-raw) As(V) [AAS-Freiberg-water-As-species]!AsV_ugL and [AAS-Freiberg-water-As-

species]!AsV_% [Query: calc-Arsenic-species] (raw data as explained for total As in table AAS-Freiberg-water-AsV-raw; ion

exchangers were eluted at least 2 times for As(V) species: the code YNPGG02-AsV1-2W and YNPGG02-AsV2-2W indicates the first and second elution of As(V) with 1 N HCl for the site YNPGG02 during the second sampling cam-paign; second, third and fourth elution of the anion exchanger with acetic acid (“MMAA” fraction; YNPGG02-MMA2-2W, YNPGG02-MMA3-2W, YNPGG02-MMA4-2W) were added to the As(V) fraction, since continuing high concentrations after the first elution peak were thought to result from desorption of As(V) already; the column AsV_final_concentration_ugL considers dilution factors during analysis; the real “final” value [AAS-Freiberg-water-As-species]!AsV_ugL is a sum of all elutions (YNPGG02-AsV1-2W + YNPGG02-AsV2-2W + YNPGG02-MMA2-2W + YNPGG02-MMA3-2W + YNPGG02-MMA4-2W) divided by 2 to correct for the enrichment factor 2 resulting from the application of 20 mL sample in the field and the elution with 10 mL extrac-tant in the laboratory)

MMAA [AAS-Freiberg-water-As-species]!MMAA_ugL and [AAS-Freiberg-water-As-species]!MMAA_% [Query: calc-Arsenic-species]

(raw data as explained for As(V) in table AAS-Freiberg-water-MMAA-raw, only one elution per sample, second, third, fourth elutions were attributed to the As(V) fraction)

DMAA [AAS-Freiberg-water-As-species]!DMAA_ugL and [AAS-Freiberg-water-As-species]!DMAA_% [Query: calc-Arsenic-species]

(raw data as explained for As(V) in table AAS-Freiberg-water-DMAA-raw; two elutions per sample)

Arsenic QC? default: invisible, only total As concentrations determined by GF-AAS are dis-played; click on the Option “Arsenic QC?” displays As additionally as deter-mined by ICP-AES and as the sum of As species from GF-AAS; the latter suffers from analytical errors of four single determinations (As(III), As(V), MMAA, DMAA;) and is thus regarded as less precise; mean, range and ROA are calcu-lated as explained in the following [Query: calc-Comparison-As]:

[Comparison-As-determination]!Sum-As-species-AAS from [AAS-Freiberg-water-As-species]! [AAS-Freiberg-water-As-species]!AsIII_ugL+[AAS-Freiberg-water-As-

species]![MMAA_ugL]+[AAS-Freiberg-water-As-species]!AsV_ugL+[AAS-Freiberg-water-As-

species]![DMAA_ugL]

[Comparison-As-determination]!Total-As-AAS [AAS-Freiberg-water-As_total]!As_final_concentration_ugL

204 Appendix

[Comparison-As-determination]!Total-As-ICP-AES [ICP-AES-water-final-data]!As_mgL⋅1000

[Comparison-As-determination]!Mean-Sum-species-AAS (analogous for Sum-species-ICP-AES, and AAS-ICP-AES) ([Comparison-As-determination]![Sum-As-species-AAS]+[Comparison-As-determination]![Total-

As-AAS])/2

[Comparison-As-determination]!Range-Sum-species-AAS (analogous for Sum-species-ICP-AES, and AAS-ICP-AES), range is not displayed in the water sub-form If(Abs([Comparison-As-determination]![Mean-Sum-species-AAS]-[Comparison-As-

determination]![Sum-As-species-AAS])>Abs([Comparison-As-determination]![Mean-Sum-

species-AAS]-[Comparison-As-determination]![Total-As-AAS]);Abs([Comparison-As-

determination]![Mean-Sum-species-AAS]-[Comparison-As-determination]![Sum-As-species-

AAS]);Abs([Comparison-As-determination]![Mean-Sum-species-AAS]-[Comparison-As-

determination]![Total-As-AAS]))

[Comparison-As-determination]!ROA-Sum-species-AAS (analogous for Sum-species-ICP-AES, and AAS-ICP-AES) [Comparison-As-determination]![Range-Sum-species-AAS]/[Comparison-As-

determination]![Mean-Sum-species-AAS])⋅100

Ion chromatography: Cl, S(VI), F, Br, N(V) [IC-water-final-data]!Cl_corrected_mgL, [IC-water-final-data]!SO4_corrected_

mgL, [IC-water-final-data]!F_corrected_mgL, [IC-water-final-data]!Br_mgL, [IC-water-final-data]!NO3_mgL;

Raw data including standard calibration from the IC are listed in table [IC-water-raw-data-area-heights]!; concentrations derived via Excel calibration curves (not shown in the Access database) are in table [IC-water-raw-data-conc]!; all sam-ples were measured in three different dilutions: undiluted, diluted 1:10, diluted 1:100, the best fit within quantification limit ([IC-QuantLimit]!) and upper cali-bration range ([IC-UpperCalibRange]!) was determined according to the follow-ing query, e.g., for F- [Query: calc-IC-water-final-values]: If([IC-water-raw-data-conc]!F_mgL_dil100>[IC-QuantLimit]!F_mgL;If([IC-water-raw-data-

conc]!F_mgL_dil100<[IC-UpperCalibRange]!F_mgL;[IC-water-raw-data-

conc]!F_mgL_dil100⋅100;999);If([IC-water-raw-data-conc]!F_mgL_dil10>[IC-

QuantLimit]!F_mgL;If([IC-water-raw-data-conc]!F_mgL_dil10<[IC-

UpperCalibRange]!F_mgL;[IC-water-raw-data-conc]!F_mgL_dil10⋅10;999);If([IC-water-raw-

data-conc]!F_mgL_dil1>[IC-QuantLimit]!F_mgL;If([IC-water-raw-data-conc]!F_mgL_dil1<[IC-

UpperCalibRange]!F_mgL;[IC-water-raw-data-conc]!F_mgL_dil1⋅1;999);If([IC-water-raw-data-

conc]!F_mgL_dil1>[IC-DetectionLimit]!F_mgL;[IC-water-raw-data-conc]!F_mgL_dil1⋅1;If([IC-

water-raw-data-conc]!F_mgL_dil1=-999;-999;-[IC-DetectionLimit]!F_mgL)))))

Appendix 205

Concentrations below the detection limit ([IC-DetectionLimit]!) are indicated as negative values; final data from the automated calculation are listed in the table [IC-water-final-data]!, e.g., [IC-water-final-data]!F_mgL; all values were checked manually afterwards, and some were corrected (e.g., [IC-water-final-data]!F_corrected_mgL), based on better fit of other dilutions or re-analyses with another IC ([IC-Freiberg-water]!), comments on each record manually adapted are given e.g., in [IC-water-final-data]!F_corrected_comment

Analysis QC? [Data-check-summary-QC]!; default: invisible, click on the Option “QC?” dis-plays the analytical error, and a comparison between measured and calculated conductivity as well as an evaluation of the analysis based on the criteria dis-cussed in section 4.2.3.1 and shown in Table 11. A general evaluation was based on a 5% criteria; whenever analytical error or range of analysis of measured and calculated conductivities were more than 5% a warning is displayed.

To determine analytical error and calculate conductivity the speciation modeling program PHREEQC was used. The query to combine data from different tables to import into PHREEQC is [PhreeqC-data-combination]. The results from PHREEQC modeling (database WATEQ4F; folder PHREEQC/ YNP03-data-check on the CD) are export via Selected_Output and imported into Access as the table [Data-check-PhreeqC-raw-data]!

To compare measured conductivity with calculated conductivities (Rossum 1975), the following is needed:

• concentrations of the cations / anions (query [calc-data-check-PhreeqC-mmolL] to convert concentrations in mol/L to mmol/L: [Data-check-PhreeqC-mmolL]!)

• the charge of the cations / anions ([Data-check-parameter-charge]!), and • their contribution to conductivity at 25°C ([Data-check-parameter-

ioncond-25]!) The following calculations are conducted:

Query [calc-data-check-cond-equiv] yields table [Data-check-cond-equiv]! [Data-check-PhreeqC-mmolL]![H_mmolL] ⋅ [Data-check-parameter-ioncond-25]![H] (analogous for the other cations / anions)

Query [calc-data-check-cZ] yields [Data-check-cZ]! [Data-check-PhreeqC-mmolL]!H_mmolL⋅ ([Data-check-parameter-charge]!H) (analogous for the other cations / anions)

Query [calc-data-check-cZ2] yields table [Data-check-cZ2]! [Data-check-PhreeqC-mmolL]!H_mmolL⋅ ([Data-check-parameter-charge]!H^2) (analogous for the other cations / anions)

And finally Query [calc-data-check-final] yields table [Data-check-final-calc]!

206 Appendix

c- Sum of anion concentrations [Data-check-PhreeqC-mmolL]!OH_mmolL+[Data-check-PhreeqC-mmolL]!CO3_mmolL+[Data-

check-PhreeqC-mmolL]!HCO3_mmolL+[Data-check-PhreeqC-mmolL]!SO4_mmolL+[Data-

check-PhreeqC-mmolL]!Cl_mmolL+[Data-check-PhreeqC-mmolL]!NO3_mmolL+[Data-check-

PhreeqC-mmolL]!F_mmolL+[Data-check-PhreeqC-mmolL]!Br_mmolL+[Data-check-PhreeqC-

mmolL]!HSO4_mmolL

c+ Sum of cation concentrations [Data-check-PhreeqC-mmolL]!H_mmolL+[Data-check-PhreeqC-mmolL]!Ca_mmolL+[Data-

check-PhreeqC-mmolL]!Mg_mmolL+[Data-check-PhreeqC-mmolL]!Na_mmolL+[Data-check-

PhreeqC-mmolL]!K_mmolL+[Data-check-PhreeqC-mmolL]!Li_mmolL+[Data-check-PhreeqC-

mmolL]!Fe2_mmolL+[Data-check-PhreeqC-mmolL]!Al_mmolL+[Data-check-PhreeqC-

mmolL]!Ba_mmolL+[Data-check-PhreeqC-mmolL]!Fe3_mmolL+[Data-check-PhreeqC-

mmolL]!Sr_mmolL

err analytical error

([Data-check-final-calc]![c+]-[Data-check-final-calc]![c-])/(0.5⋅([Data-check-final-

calc]![c+]+[Data-check-final-calc]![c-]))

G0- sum of concentration ⋅ equivalent conductance for anions

[Data-check-cond-equiv]!OH_mmolL+[Data-check-cond-equiv]!CO3_mmolL+[Data-check-cond-

equiv]!HCO3_mmolL+[Data-check-cond-equiv]!SO4_mmolL+[Data-check-cond-

equiv]!Cl_mmolL+[Data-check-cond-equiv]!NO3_mmolL+[Data-check-cond-

equiv]!F_mmolL+[Data-check-cond-equiv]!Br_mmolL+[Data-check-cond-equiv]!HSO4_mmolL

G0+ sum of concentration ⋅ equivalent conductance for cations

[Data-check-cond-equiv]!H_mmolL+[Data-check-cond-equiv]!Ca_mmolL+[Data-check-cond-

equiv]!Mg_mmolL+[Data-check-cond-equiv]!Na_mmolL+[Data-check-cond-

equiv]!K_mmolL+[Data-check-cond-equiv]!Li_mmolL+[Data-check-cond-

equiv]!Fe2_mmolL+[Data-check-cond-equiv]!Al_mmolL+[Data-check-cond-

equiv]!Ba_mmolL+[Data-check-cond-equiv]!Fe3_mmolL+[Data-check-cond-equiv]!Sr_mmolL

cZ2- sum of concentrations ⋅ square of charge for anions

[Data-check-cZ2]!OH_mmolL+[Data-check-cZ2]!CO3_mmolL+[Data-check-

cZ2]!HCO3_mmolL+[Data-check-cZ2]!SO4_mmolL+[Data-check-cZ2]!Cl_mmolL+[Data-check-

cZ2]!NO3_mmolL+[Data-check-cZ2]!F_mmolL+[Data-check-cZ2]!Br_mmolL+[Data-check-

cZ2]!HSO4_mmol

cZ2+ sum of concentrations ⋅ square of charge for anions

[Data-check-cZ2]!H_mmolL+[Data-check-cZ2]!Ca_mmolL+[Data-check-

cZ2]!Mg_mmolL+[Data-check-cZ2]!Na_mmolL+[Data-check-cZ2]!K_mmolL+[Data-check-

cZ2]!Li_mmolL+[Data-check-cZ2]!Fe2_mmolL+[Data-check-cZ2]!Al_mmolL+[Data-check-

cZ2]!Ba_mmolL+[Data-check-cZ2]!Fe3_mmolL+[Data-check-cZ2]!Sr_mmolL

cZ- sum of products concentrations ⋅ charge for anions

Appendix 207

[Data-check-cZ]!OH_mmolL+[Data-check-cZ]!CO3_mmolL+[Data-check-

cZ]!HCO3_mmolL+[Data-check-cZ]!SO4_mmolL+[Data-check-cZ]!Cl_mmolL+[Data-check-

cZ]!NO3_mmolL+[Data-check-cZ]!F_mmolL+[Data-check-cZ]!Br_mmolL+[Data-check-

cZ]!HSO4_mmolL

cZ+ sum of products concentrations ⋅ charge for cations

[Data-check-cZ]!H_mmolL+[Data-check-cZ]!Ca_mmolL+[Data-check-cZ]!Mg_mmolL+[Data-

check-cZ]!Na_mmolL+[Data-check-cZ]!K_mmolL+[Data-check-cZ]!Li_mmolL+[Data-check-

cZ]!Fe2_mmolL+[Data-check-cZ]!Al_mmolL+[Data-check-cZ]!Ba_mmolL+[Data-check-

cZ]!Fe3_mmolL+[Data-check-cZ]!Sr_mmolL

Z- effective ion charge anions

[Data-check-final-calc]![cZ2-]/[Data-check-final-calc]![cZ-]

Z+ effective ion charge cations

[Data-check-final-calc]![cZ2+]/[Data-check-final-calc]![cZ+]

Lambda- average value for ionic conductance of the anions

[Data-check-final-calc]![G0-]/[Data-check-final-calc]![c-]

Lambda+ average value for ionic conductance of the cations

[Data-check-final-calc]![G0+]/[Data-check-final-calc]![c+]

Lambda 0 total ionic conductance

[Data-check-final-calc]![lambda-]+[Data-check-final-calc]![lambda+]

Q

([Data-check-final-calc]![Z+]⋅[Data-check-final-calc]![Z-]⋅[Data-check-final-

calc]!lambda0)/(([Data-check-final-calc]![Z+]+[Data-check-final-calc]![Z-])⋅([Data-check-final-

calc]![Z+]⋅[Data-check-final-calc]![lambda-]+[Data-check-final-calc]![Z-]⋅[Data-check-final-

calc]![lambda+]))

C

([Data-check-final-calc]![c-]+[Data-check-final-calc]![c+])/2

G calculated conductivity

[Data-check-final-calc]![G0+]+[Data-check-final-calc]![G0-]-(((([Data-check-final-calc]!lambda0⋅

[Data-check-final-calc]![Z+]⋅[Data-check-final-calc]![Z-])/(115.2⋅([Data-check-final-

calc]![Z+]+[Data-check-final-calc]![Z-])))⋅((2⋅[Data-check-final-calc]!Q)/(1+[Data-check-final-

calc]!Q^0.5)))+0.668⋅([Data-check-final-calc]![Z+]+[Data-check-final-calc]![Z-])⋅[Data-check-

final-calc]!C)^1.5

Finally, [Data-check-summary-QC]! lists the percentage error, mean and ROA of measured and calculated conductivity, as well as a comment on the reason for deviations ([Data-check-summary-QC]!Comment) and an evaluation of the analysis ([Data-check-summary-QC]!Comment2) according to the following query [calc-data-check-Summary-QC]:

Comment: If([Data-check-summary-QC]!percent_error<0 and [Data-check-summary-

QC]!conductivity_calculated>[Data-check-summary-QC]!conductivity_measured;"too many ani-

208 Appendix

ons";If([Data-check-summary-QC]!percent_error<0 and [Data-check-summary-

QC]!conductivity_calculated<[Data-check-summary-QC]!conductivity_measured;"not enough

cations";If([Data-check-summary-QC]!percent_error>0 and [Data-check-summary-

QC]!conductivity_calculated>[Data-check-summary-QC]!conductivity_measured;"too many

cations";"not enough anions")))

Evaluation: If([Data-check-summary-QC]!percent_error>5 and [Data-check-summary-QC]![ROA%]>5; “At-

tention! Analysis off by more than 5%!”; If([Data-check-summary-QC]!percent_error>5 and

[Data-check-summary-QC]![ROA%]<5; “Attention! Analytical error too high, even though con-

ductivity seems ok”; If([Data-check-summary-QC]!percent_error<-5 and [Data-check-summary-

QC]![ROA%]>5; “Attention! Analysis off by more than 5%!”; If([Data-check-summary-

QC]!percent_error<-5 and [Data-check-summary-QC]![ROA%]<5; “Attention! Analytical error

too high, even though conductivity seems ok”; If([Data-check-summary-QC]!percent_error<5 and

[Data-check-summary-QC]![ROA%]>5; “Attention! Range of analysis for conductivities >5%,

even though analytical error seems ok”; “analysis ok”)))))

Carbon: TOC [TOC-water-mean]!TOC_mean_mgL; value is calculated as the mean of double

determinations ([TOC-water-raw-data]!TOC_1st-determ_mgL, and [TOC-water-raw-data]!TOC_2nd-determ_mgL) [Query: calc-TOC-mean]

TIC [TIC-water-mean]!TIC_mean_mgL; value is calculated as the mean of double determinations ([TIC-water-raw-data]!TIC_1st-determ_mgL, and [TIC-water-raw-data]!TIC_2nd-determ_mgL) [Query: calc-TIC-mean]

Mercury: [Mercury]!Hg_ngL, [Mercury]!MeHg_ngL [Mercury]!Sampling_date, [Mer-

cury]!Sampling_time

Appendix 209

Mercury data only exist for 12 samples, thus these data had to be attached via a subform [Form Mercury]; in the main form data fields will be empty where no data exist

ICP-AES: As, Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Si, Se, Sr, V, Zn

[ICP-AES-water-final-data]! Raw data from the ICP-AES are listed in table [ICP-AES-water-raw-data]!; in-

formation on Wavelength, Scan Mode, and ControlMaterial given in the tables [ICP-AES-Wavelength]!, [ICP-AES-ScanMode]!, [ICP-MS-ControlMaterial]!); all samples were measured in three different dilutions: undiluted, diluted 1:10, diluted 1:100, the best fit within quantification limit ([ICP-AES-QuantLimit]!) and upper calibration range ([ICP-AES-UpperCalibRange]!) was determined via a query [Query: calc-ICP-AES-water-final-values] analogous to shown before for IC. Concentrations below the detection limit ([ICP-AES-DetectionLimit]!) are indicated as negative values; like for IC all values were checked manually af-terwards, and some were corrected

ICP-MS: Li, Be, Na, Mg, Al, Si, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr, Y, Zr, Nb, Mo, Ru, Pd, Ag, Cd, In, Sn, Sb, Te, I, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Ta, W, Re, Os, Pt, Au, Hg, Tl, Pb, Bi, Th, U

[ICP-MS-water]! ICP-MS data only exist for 27 samples, thus these data had to be attached via a

subform [Form ICP-MS-water]; in the main form data fields will be empty where no data exist

Site Code at the bottom of the water subform page is only displayed for better orientation (it also essentially connects the water subform to the main data form, so that not all 62 records are displayed for all sites, but only the records with matching site codes)

Third page: Gas - total concentrations The page “Gas - total conc.” contains data from the subforms “Gas-NaOCl-subform” (75 records), and “Gas-NaOH-subform” (48 records).

Gas-NaOCl subform (NaOCl samples): primary key Analysis-code-gas-NaOCl ([Analysis-code-gas-NaOCl]!Analysis-code-gas-

NaOCl); the analysis code gas NaOCl consists of the site code, the No. of the sampling campaign (1 = June, 5th, to June, 21st, 2003; 2 = June, 30th, to July, 12th, 2003; 3 = September, 3rd, to September, 15th, 2003; 4 = September 21st, to Octo-ber, 9th, 2003), and G for gas, NaOCl for the oxidizing reagent, and an abbrevia-

210 Appendix

tion for gas hood – oxidizing bottle –setup (according to Figure 42: wGH = with gas hood, wGH-tni = with gas hood, tube not immersed, nGH = no gas hood, nGH-cv = no gas hood, closed valves)

Sampling from… to [Analysis-code-gas-NaOCl]!Date_start, Time_start [Analysis-code-gas-NaOCl]!Date_end, Time_end Duration [NaOCl-time-calc]!Duration_exposed_days,

[NaOCl-time-calc]!Duration_exposed_hours [Query: calc-NaOCl-time]

The table [Analysis-code-gas-NaOCl]! as well as the other [analysis-code-gas]! tables also contain information about type of gas hood and oxidizing bottle used in each experiment that is not dis-played in the respective subforms ([Analysis-code-gas-NaOCl]!Gashood_type from table: [Gashood_Type]! and ([Analysis-code-gas-NaOCl]!Oxidation_bottle_type from table [Oxida-tion_bottle_type!] with information on material, and size). Figure in the center shows the setup type according to Figure 42 (link in [Pictures-Gashood-

Setup]!Picture1; pictures in folder Figures / Database-Yellowstone2003) Concentrations given on the left side are underneath the GH “wGH” (fields are empty, when sam-pling was done without gas hood), concentrations on the right side are in ambient air (“nGH”). Calculation from dissolved concentrations in the oxidizing solution to concentrations in the gas phase are done according to the following equation (details explained in section 4.2.3.2):

DAt∆xm c⋅⋅

⋅=

Appendix 211

The origin of the single parameters is explained in the following for the case with gas hood , the non gas hood case goes analogous. Note: As described in section 4.2.3.2 concentrations for setups with gas hoods are upper limits for the two end-member models of zero concentration in ambient air and underneath the gas hood (a similar warning is displayed at the end of the page “Gas - total conc.”).

AAS (As) m conc. [µg] = [AAS-Boulder-gas-NaOCl-considered-data]! [µg/L] ⋅ 0.1 [L] oxidizing solution * raw data from up to 3 triple determinations in [AAS-Boulder-gas-NaOCl-raw]! * converted to final data calculating means of the triple determinations and con-sidering dilution factor ([AAS-Boulder-water-final]! via Query: calc-AAS-Boulder-gas-NaOCl) * final data are corrected for the max value of three blank determinations (2 lab blanks 1:100 diluted NaOCl (YNPBW02-1G-NaOCl, YNPBW04-2G-NaOCl), and 1 field blank YNPHL01-3G-NaOCl-nGH-cv in table [AAS-Boulder-gas-field-lab-blanks]!), [AAS-Boulder-gas-blank-max]! contains the maximum As blank; Query [calc-AAS-Boulder-gas-NaOCl-considered-data] corrects every As value for this blank and writes it back to table [AAS-Boulder-gas-NaOCl-considered-data]! considering values below detection limit ([AAS-DL-QL-UCR]!) as negative values ∆x [Gas-Difflength-area-NaOCl-NaOH]!wGH_Diffusion_length_cm A [Gas-Difflength-area-NaOCl-NaOH]!wGH_Area_cm2 t [NaOCl-time-calc]!Duration_exposed_hours ⋅ 60 ⋅ 60 D [Gas-Diffcoeff-NaOCl-As]! Diffusion_coefficient_mean_cm2s-1 (Query [calc-Gas-Diffcoeff-NaOCl-As] calculated as explained in section 4.2.3.2 ac-cording to the following equation:

2

0.51.75

)V(V P

MT0.001 D

1/31/3BA

r

−⋅

⋅= )

Option “show range” (default invisible; click displays minimum and maximum values considering

uncertainties for m, ∆x, and D; uncertainties of A and t were considered to be negligible) m max. concentration: m ⋅ 1.01; min. concentration: m = 0.99 ∆x [Gas-Difflength-area-NaOCl-NaOH]!wGH_max_Diffusion_length_cm considering wGH_Diffusion_length_uncertainty_cm via Query [calc-Gas-Difflength-area-NaOCl-NaOH], analogous for minimum length D maximum diffusion coefficient considers uncertainty in temperature, and dominant species in gas, maximum value is for Temperature +10°C and an as-sumed predominance of AsH3 (with the lowest molecular weight for any volatile species), minimum values is for Temperature -10°C and an assumed predomi-

212 Appendix

nance of AsCl3 (with the highest molecular weight) according to the following formulas in Query [calc-Gas-Diffcoeff-NaOCl-As]: 0.001⋅[Gas-Diffcoeff-NaOCl-As]![Temperature_at_30cm_K+10]^1.75⋅[Gas-Diffcoeff-NaOCl-

As]!Mr_AsH3^0.5/[Gas-Diffcoeff-NaOCl-As]!p_bar/([Gas-Diffcoeff-NaOCl-As]![VA_cm3mol-

1]^(1/3)+[Gas-Diffcoeff-NaOCl-As]![VB_AsH3_cm3mol-1]^(1/3))^2

0.001⋅[Gas-Diffcoeff-NaOCl-As]![Temperature_at_30cm_K-10]^1.75⋅[Gas-Diffcoeff-NaOCl-

As]!Mr_AsCl3^0.5/[Gas-Diffcoeff-NaOCl-As]!p_bar/([Gas-Diffcoeff-NaOCl-As]![VA_cm3mol-

1]^(1/3)+[Gas-Diffcoeff-NaOCl-As]![VB_AsCl3_cm3mol-1]^(1/3))^2

Since the diffusion coefficient shows an inverse correlation with concentration in the gas phase, minimum diffusion coefficients had to be taken for maximum concentrations and vice versa-

ICP-AES (analogous to HG-AAS explained before) m [ICP-AES-gas-NaOCl-raw-data]!; [ICP-AES-gas-NaOCl-final-data]! Query: calc-ICP-AES-gas-NaOCl-final-values considers the different dilutions as explained for ICP-AES water before; values in [ICP-AES-gas-NaOCl-considered-data]! are corrected for the maximum blank values ([ICP-AES-gas-field-lab-blanks]! and [ICP-AES-gas-blank-max]!); only elements with concen-trations significantly above detection limit and blank value were considered: As, Al, B, Ba, Cu, Fe, K, Li, SiO2, Sr, Zn ∆x, A, t, and D are the same values as for As determination described before

Range for m, uncertainties of ± 2% were considered, expect for As (± 1%) and K (± 3%)

IC (S(VI)) (analogous to HG-AAS explained before) m [IC-Freiberg-gas-NaOCl-SO4-2]!SO4_mgL; uncertainty ± 5%

ICP-MS (analogous to HG-AAS and ICP-AES explained before) m [ICP-MS-gas-NaOCl-raw-data]!; [ICP-MS-gas-NaOCl-final-data]!: Query: calc-ICP-MS-gas-NaOCl-final-values considers dilution of NaOCl sam-ples before sending them to Actlab (1:1); [ICP-MS-gas-NaOCl-considered-data]! corrected for blank value ([ICP-MS-gas-blank]!), and only elements significantly above detection limit ([ICP-MS-Detection-Limit]!) and blank: Li, Mg, Al, Si, K, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Rb, Sr, Y, Zr, Nb, Mo, Cd, Sb, Cs, Ba, La, Ce, Pr, Nd, Sm, Gd, Dy, Er, Yb, Hf, W, Pb, Th.

Range is not displayed for ICP-MS due to the large amount of data disturbing the form´s clarity. If required it can easily be calculated the same way as described for ICP-AES.

Analysis code at the end of the page displayed again just for orientation

Appendix 213

Gas-NaOH-subform (NaOH samples) primary key Analysis-code-gas-NaOH [Analysis-code-gas-NaOH]!Analysis-code-gas-

NaOH; analysis code analogous to analysis code NaOCl described before Sampling from… to, Duration analogous to fields described before for NaOCl Figure in the center shows the setup type according to Figure 42 (link in [Pictures-Gashood-

Setup]!Picture2; pictures in folder Figures / Database-Yellowstone2003) Photometry (Cl)

m [Photometry-gas-NaOH-Cl]!Cl-_mgL Gas-Diffcoeff-NaOH-Cl; uncertainty ± 5% D [Gas-Diffcoeff-NaOH-Cl]!Diffusion_coefficient_mean_cm2s-1 (for range only temperature uncertainty was considered with ± 10°C) (Query: [calc-Gas-Diffcoeff-NaOH-Cl])

Fourth page: Gas comparison wGH-nGH The page “Gas comparison wGH-nGH” contains data from the subforms “Comparison-wGH-nGH-NaOCl-As” (17 records), and “Comparison-wGH-nGH-NaOH” (8 records). While the concentra-tions for gas hood setups (wGH) on page “Gas- total conc.” are only upper limits for either ambient air or gas hood, this page takes concentrations in ambient air from nGH setups as representative for wGH setups, too, and considering this concentration, calculates the mean concentration underneath the gas hood as described in section 4.2.3.2. Since this was only possible at sites, where nGH and wGH were done parallel the dataset is smaller.

214 Appendix

Subform “Comparison-wGH-nGH-NaOCl-As” Site codes of setup with (wGH on top) and without gas hood (nGH below); primary key is

nGH for calculations in ambient air, for all others it is wGH AAS ambient air: [Gas-ugm-3-corrected-NaOCl-wGH-nGH-mean]!nGH_As_ugm-3

([essentially the same as Gas-ugm-3-NaOCl-mean]!nGH_As_ugm-3, only nega-tive values (below detection limit) were replaced by -0.3 ⋅ detection limit, since this value (c1) is needed for the following calculation of c2 according to equation:

)∆x∆c

DAtm(∆x∆c

1

122 −

⋅⋅⋅=

under gas hood [Gas-ugm-3-corrected-NaOCl-wGH-nGH-mean]!wGH_As_ugm-3 (calculated via the Query [calc-Gas-ugm-3-corrected-NaOCl-wGH-nGH-mean])

decrease to… [%] ambient air : gas hood ⋅ 100 in table [Gas-ugm-3-corrected-NaOCl-wGH-nGH-gradientsAs]!mean-gradient-wGH-nGH from Query [calc-Gas-ugm-3-corrected-NaOCl-gradientsAs]

Range calculations done as explained before. For a maximum value of c2, maximum x2, maximum m, and minimum D, c1, and x1 were taken, vice versa for minimum c2. Only As concentrations from HG-AAS are displayed in this main data form. The tables [calc-Gas-ugm-3-corrected-NaOCl-wGH-nGH-mean]!, max and min also contain the data for all other ele-ments determined by ICP-AES and for S(VI) determined by IC. Subform “Comparison-wGH-nGH-NaOH”

Analogous to As explained before. Table [Gas-mgm-3-corrected-NaOH]! Contains mean, maximum, and minimum concentrations for ambient air and gas hood derived from Query [calc-Gas-mgm-3-corrected-NaOH]

Appendix 215

Fifth page: Gas - total conc. (P, DB) The page “Gas – total conc. (P, DB)” contains data from the subforms “Gas-NaOCl-P-subform” (4 records), and “Gas-NaOCl-DB-subform” (21 records).

Gas-NaOCl-P-subform primary key Analysis-code-gas-NaOCl-P [Analysis-code-gas-NaOCl-P]! Analysis-code-

gas-NaOCl-P; analysis code analogous to analysis code NaOCl described before Sampling from… to, Duration analogous to fields described before for NaOCl Sampled gas volume [NaOCl-P-time-amount-calc]! sampled_gas_volume_L calculated from

pump_hubs ⋅ 0.1 L per hub via Query [calc-NaOCl-P-time-amount]

AAS, ICP-AES Concentration is calculated as Vm c = as described in section 4.2.3.2

For m the tables [AAS-Boulder-gas-NaOCl-considered-data]! and [ICP-AES-gas-NaOCl-considered-data]! were used; results in [Gas-ugm-3-NaOCl-P] (Query [calc-Gas-ugm-3-NaOCl-P])

Gas-NaOCl-DB-subform primary key Analysis-code-gas-NaOCl-DB [Analysis-code-gas-NaOCl-DB]! Analysis-code-

gas-NaOCl-DB; analysis code analogous to analysis codes described before Sampling from… to, Duration analogous to fields described before for NaOCl AAS Only As was determined for DB setups [AAS-Boulder-gas-NaOCl-considered-

data]!; results in [Gas-ugm-3-NaOCl-DB]!As_ugm-3_mean, max, and min from Query [calc-Gas-ugm3-NaOCl-DB] as described before on page “gas – total conc.”

216 Appendix

Sixth page: Gas - As speciation The page “Gas - As speciation” contains records from the “SPME-subform” (60 records) in the center as well as a link to the “SS-subform” (40 records) in the lower right corner.

SPME-subform (solid-phase micro extraction) primary key [Analysis-code-gas-SPME]!Analysis-code-gas-SPME; analysis code analogous

to analysis code NaOCl described before, instead of NaOCl type of fiber (PDMS, CAR, DVB) and No. of GC-MS analysis campaign (1, 2, 4 analyzed in Denver, 3 and 5 in Bozeman); respective GC-MS equipment, columns, and op-erating conditions are in table [GC-MS-conditions]!

Sampling from… to, Duration analogous to fields described before SPME type [Analysis-code-gas-SPME]!SPME_type, detailed description of types, applica-

tion range, max. temperatures, pH conditions, etc. in table [SPME-Type]! Picture1 Chromatogram; link in [Pictures-Chromatograms]!Picture1, pictures saved in

Folder Figures / Database-Yellowstone2003; double-click on the small picture opens an enlarged version of the same picture ([Pictures-Chromatograms]!Picture1B)

Picture2 detected (CH3)3As; [Pictures-Chromatograms]!Picture2 and Picture2B. If there is no volatile (CH3)3As detected in the chromatogram, a picture with the text “No volatile As species found in chromatogram” is displayed.

Picture3 (CH3)2AsCl Picture4 (CH3)AsCl2 Picture5 (CH3)2AsS(CH3)

Appendix 217

Area [SPME-peak-area]! Retention time [SPME-retention-time]!

The subform for solid sorbents (SS-subform) is not included in the main data form, since no vola-tile As species were found on solid sorbents at all. Via the command “Open Subform Solid Sor-bents” in the lower right corner it is possible to get general information on solid sorbent setup for each site. The sub-window is blank if no solid sorbent was used for the site. Information on sam-pling date and time origin from the respective tables as explained before. Primary key is [Analysis-code-gas-SS]!Analysis-code-gas-SS. The table [SS-Type]! contains detailed description of SS types, surface area, sampling range, max. temperatures, etc.

218 Appendix

Appendix 10 Overview of sampling sites with parameters determined in the field campaigns from June, 5th to July, 16th, 2003 and from September, 2nd to October, 12th 2003

Are

a

Site

Stan

dard

w

ater

sa

mpl

e1

Wat

er

sam

ple

ICP-

MS2

Mer

cury

3

Gas

sam

ple

NaO

Cl4

Gas

sam

ple

NaO

Cl

ICP-

MS5

Gas

sam

ple

P / D

B6

Gas

sam

ple

NaO

H7

Gas

sam

ple

SS8

Gas

sam

ple

SPM

E9

YNPGG01 05.07.S1 08.07.S1 12.07.

YNPGG02 03.07. 05.07.S1 02.10.S1,3 10.07.S1

07.10.S1,3 12.07. CAR

YNPGG03 05.07.S1 02.10.S1,3 10.07.S1

07.10.S1,3 12.07. CAR

YNPGG04 03.07. 11.07. 22.09.

22.09. 05.07.S1 02.10.S1,3 02.10.S1 10.07.

P 08.07.S1 07.10.S1,3 12.07.

YNPGG05 10.07. 07.10. 05.07.S1

02.10.S1,3 08.07.S1 07.10.S1,3 12.07.

YNPGG06 05.07.

YNPGG07 10.07. 07.10. 02.10.S1 07.10.S1

Gib

bon

Gey

ser B

asin

YNPGG08 07.10. 02.10.S3 07.10.S3

YNPHL01 18.06. 09.07. 09.09.

18.06. 09.09. 07.06.

18.06.S1 04.07.S1

12.09.S1,3,4

18.06.S1 12.09.S1

09.07.P

09.09.DB

09.07.S1 15.09.S1 30.06.

21.06. CAR 12.07. CAR, DVB

15.09. CAR

YNPHL02 16.06. 07.06. 18.06.S1 30.06.

Haz

le L

ake

YNPHL03 14.09. 15.09.S3 09.09.DB 15.09.S3 15.09. CAR

YNPLG01 01.07. 01.07. 17.06.S1 09.07.S1 09.07.S1 12.07.S1

YNPLG02 02.07. 02.07. 05.07.S1 05.07.S1 09.07.S1 12.07. 12.07. CAR, PDMS

Low

er G

eyse

r B.

YNPLG03 02.07. 02.07. 06.07.S1 06.07.S1 09.07.S1 12.07.

YNPNL01 12.06. 18.06.S1

YNPNL02 12.06. 06.07. 14.09.

12.06. 18.06. MeHg

**18.06.S1

12.07.S3 04.07.S1 *12.09.S1

18.06.S1 09.09.DB

08.07.S1 12.09.S1 12.07.

21.06. CAR 12.07. CAR

09.09. CAR, DVB

YNPNL03 18.06. 06.07. 14.09.

18.06. 18.06.S1 06.07.S1

**12.09.S118.06.S1 08.07.S1

12.09.S1 12.07. 09.09. CAR, DVB

YNPNL04 12.06. 07.07. 13.09.

12.06. 18.06. 18.06.S1 07.07.S1

*12.09.S1 18.06.S1 09.09.

DB 30.06.S1 12.09.S1 30.06.

21.06. CAR, DVB, PDMS

09.09. CAR, DVB, PDMS

YNPNL05 16.06. 07.07. 16.06. 18.06. 20.06.S1

07.07.S1 20.06.S1 30.06.S1 30.06. Nym

ph L

ake

/ Fry

ing

Pan

Sprin

g

YNPNL06 20.06.S1

YNPRH01 11.06. 04.07. 14.09.

11.06. 14.09. 17.06.

17.06.S1 04.07.S1

15.09.S1,2,3

17.06.S1 15.09.S1

11.07.P

08.09.DB

30.06.S1 11.07.S1 30.06. 21.06. CAR, PDMS

09.09. CAR, PDMS

YNPRH02 17.06.S1 30.06.S1 30.06. 21.06. CAR YNPRH03 11.06. 17.06. 17.06.S1 04.07.S1 30.06. 21.06. CAR

YNPRH04 19.06. 04.07. 14.09.

19.06. 14.09.

04.07. MeHg

20.06.S1 04.07.S1 15.09.S1,3

20.06.S1 15.09.S1

30.06.S1 08.07.S1 15.09.S3

11.07. 09.09. CAR, PDMS15.09. DVB

Rag

ged

Hill

s

YNPRH05 14.06. 13.09.

14.06. 13.09.

17.06.S1 15.09.S1,3 30.09.S12,3

17.06.S1 15.09.S1 30.06.S1

15.09.S1 30.06.

21.06. CAR 09.09. CAR, DVB 15.09. DVB (2x),

PDMS 07.10. CAR, PDMS

Appendix 219

Appendix 10 A

rea

Site

Stan

dard

w

ater

sa

mpl

e1

Wat

er

sam

ple

ICP-

MS2

Mer

cury

3

Gas

sam

ple

NaO

Cl4

Gas

sam

ple

NaO

Cl

ICP-

MS5

Gas

sam

ple

P / D

B6

Gas

sam

ple

NaO

H7

Gas

sam

ple

SS8

Gas

sam

ple

SPM

E9

YNPRH06

20.06.S1 11.07.S3 04.07.S1 15.09.S3

20.06.S1

15.09.S3

11.07. P 08.07.S1 30.06.

11.07. CAR 09.09. CAR 15.09. CAR, DVB, PDMS 07.10. CAR,

DVB

YNPRH07 11.07. 11.07. 20.06.S1 04.07.S1 04.07.S1 30.06.S1

08.07.S1 30.06. 11.07. CAR, PDMS

YNPRH08 15.06. 29.09. 29.09. 17.06. 20.06.S1

08.07.S1 30.06.S1 30.06.

YNPRH09 15.06. 17.06.MeHg 20.06.S1 30.06.

YNPRH10 29.09. 29.09. 15.09.S3 15.09.S3 08.09.DB 15.09. CAR

YNPRH11 12.09. 15.09.S3 30.09.S1,3

**30.09.S2

15.09.S3 15.09. CAR,

PDMS

YNPRH12 15.09.S3 30.09.S1 15.09. CAR

YNPRH13 13.09. 13.09. YNPRH14 13.09. YNPRH15 13.09. 07.10. CAR YNPRH16 29.09. YNPRH17 29.09. 29.09. YNPRH18 30.09. 30.09. 01.06. YNPRH19 30.09. 30.09. YNPRH20 30.09. 30.09. 30.05. 05.10.S1 05.10.S1 09.10.S1 YNPRH21 02.10. 02.10. YNPRH22 02.10. 05.10.S1,2,3 09.10.S1,2,3

YNPRH23 05.10.S1,2

**05.10.S3 09.10.S1,2,3

YNPRH24 02.10. YNPRH25 02.10.

YNPRH26 05.10. 05.10.

07.10. DVB

(3x), CAR (3x), PDMS (2x)

YNPRH27 05.10. 05.10.

Rag

ged

Hill

s, N

orris

Gey

ser B

asin

YNPRH28 05.10. Σ 48 60 27 12 80 21 49 20 60

1 Standard water sample includes measurement of temperature, pH, conductivity, on-site photometry on Fe(II)/Fe(tot),

S(-II), SiO2, total As, As species (As(III) / As(V) / MMAA / DMAA), IC (Cl, S(VI), F, Br, N(V)), TOC, TIC, ICP-AES (As, Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Si, Se, Sr, V, Zn); redox potential and oxygen were not determined at all sites and all times due to unstable values and problems with the devices due

2 Water sample ICP-MS was analyzed for Li, Be, Na, Mg, Al, Si, K, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Ga, Ge, As, Se, Br, Rb, Sr, Y, Zr, Nb, Mo, Ru, Pd, Ag, Cd, In, Sn, Sb, Te, I, Cs, Ba, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, Lu, Hf, Ta, W, Re, Os, Pt, Au, Hg, Tl, Pb, Bi, Th, U

3 Mercury sampling includes inorganic Hg and Methyl-Hg unless otherwise stated 4 Gas sample NaOCl was analyzed for the sum of volatile metallics by ICP-AES, for As by HG-AAS; and for dissolved

S(VI) from volatile H2S by IC; S1-4 = Setup No. according to Figure 42; * not enough sample left for ICP-AES analy-sis, ** samples analyzed, but results dismissed due to possible contaminations in the field (bottles not tight)

5 Gas sample NaOCl ICP-MS was analyzed for the sum of volatile metallics by ICP-MS 6 Gas sample NaOCl collected by active pumping (P) and by diffusion bowls (DB)

7 Gas sample NaOH was analyzed for dissolved Cl from volatile HCl 8 Gas sample SS = solid sorbent (Tenax TA and Carboxen 564) 9 Gas sample SPME = Solid-phase micro-extraction fibers (PDMS = polydimethylsiloxane, PDMS-CAR = polydimethyl-

siloxane Carboxen, PDMS-CAR-DVB = polydimethylsiloxane Carboxen divinylbenzene)

220 Appendix

Appendix 11 Dendrogram of cluster analysis of 60 water samples by Ward´s method (program: SPSS for Win-dows 11.0)

* Manual adaptations made to the membership in this dendrogram for the final cluster membership are: YNPRH07-2W shifted from subgroup 5 to subgroup 1 (as it is in Average Linkage), YNPHL03-3W shifted from subgroup 7 to subgroup 1 (as it is in Average Linkage), YNPRH25-4W, YNPRH26-4W joined to subgroup 13, subgroup 14 separated from sub-group 13

Appendix 221

Appendix 12 Dendrogram of cluster analysis of 60 water samples by the Average Linkage Within Groups method (program: SPSS for Windows 11.0)

222 Appendix

Appendix 13 Kruskall-Wallis test for overall difference between the 14 subgroups derived from cluster analysis for water samples after Ward´s method, Average Linkage Within Groups method and an optimized cluster analy-sis based on Ward´s method with manual adaptation as explained in Appendix 11; df = degree of freedom; signifi-cance levels > 1% (not accepted as significant) are marked in bold

Ward´s method Average Linkage Within

Group method Optimized cluster analysis

(Ward and manually adapted)

Chi

-Squ

are

df

sign

ifica

nce

leve

l

Chi

-Squ

are

df

sign

ifica

nce

leve

l

Chi

-Squ

are

df

sign

ifica

nce

leve

l

{H+} 46 13 0.00% 40 13 0.01% 48 13 0.00% temperature 41 13 0.01% 36 13 0.07% 40 13 0.01% conductivity 49 13 0.00% 51 13 0.00% 49 13 0.00%

F 50 13 0.00% 48 13 0.00% 52 13 0.00% Cl 54 13 0.00% 53 13 0.00% 55 13 0.00% Br 54 13 0.00% 53 13 0.00% 56 13 0.00%

N(V) 25 13 2.40% 21 13 7.70% 23 13 4.20% S(VI) 50 13 0.00% 50 13 0.00% 50 13 0.00%

As(III) 53 13 0.00% 51 13 0.00% 54 13 0.00% As(V) 51 13 0.00% 50 13 0.00% 50 13 0.00%

MMAA 46 13 0.00% 43 13 0.00% 44 13 0.00% DMAA 49 13 0.00% 39 13 0.02% 50 13 0.00%

Al 44 13 0.00% 42 13 0.01% 44 13 0.00% B 53 13 0.00% 52 13 0.00% 54 13 0.00% Ba 40 13 0.01% 30 13 0.51% 49 13 0.00% Be 44 13 0.00% 36 13 0.07% 41 13 0.01% Ca 39 13 0.02% 27 13 1.20% 42 13 0.01% Cd 28 13 0.93% 31 13 0.35% 29 13 0.66% Co 30 13 0.52% 29 13 0.71% 29 13 0.61% Cr 46 13 0.00% 50 13 0.00% 46 13 0.00% Cu 22 13 5.00% 19 13 11.00% 14 13 38.00% K 43 13 0.01% 36 13 0.05% 40 13 0.01% Li 53 13 0.00% 51 13 0.00% 56 13 0.00% Mg 45 13 0.00% 39 13 0.02% 43 13 0.00% Mn 46 13 0.00% 39 13 0.02% 44 13 0.00% Mo 52 13 0.00% 50 13 0.00% 51 13 0.00% Na 52 13 0.00% 51 13 0.00% 55 13 0.00% Ni 40 13 0.01% 38 13 0.03% 41 13 0.01% Pb 28 13 1.00% 38 13 0.03% 25 13 2.10% Sr 41 13 0.01% 29 13 0.69% 40 13 0.02% V 59 13 0.00% 59 13 0.00% 59 13 0.00% Zn 48 13 0.00% 42 13 0.01% 45 13 0.00%

SiO2 51 13 0.00% 48 13 0.00% 51 13 0.00% S(-II) 40 13 0.02% 32 13 0.21% 35 13 0.10% Fe(II) 47 13 0.00% 45 13 0.00% 46 13 0.00% Fe(III) 42 13 0.01% 36 13 0.06% 42 13 0.01%

TIC 41 13 0.01% 30 13 0.49% 38 13 0.03% TOC 40 13 0.01% 40 13 0.01% 43 13 0.00%

Appendix 223

Appendix 14 Bivariate correlation for water sample parameters (first line = Spearman correlation coefficient, second line = two tailed level of significance in %; for all parameters number of cases n = 60)

{H+} temp cond F Cl Br N(V) S(VI) As As(III) As(V) MMAA DMAA

{H+} 1.00 --- --- --- --- --- --- --- --- --- --- --- --- . --- --- --- --- --- --- --- --- --- --- --- ---

temp -0.45 1.00 --- --- --- --- --- --- --- --- --- --- --- 0.0% . --- --- --- --- --- --- --- --- --- --- ---

cond 0.09 -0.01 1.00 --- --- --- --- --- --- --- --- --- --- 49.3% 91.6% . --- --- --- --- --- --- --- --- --- --- F -0.67 0.50 0.50 1.00 --- --- --- --- --- --- --- --- --- 0.0% 0.0% 0.0% . --- --- --- --- --- --- --- --- ---

Cl -0.58 0.30 0.66 0.81 1.00 --- --- --- --- --- --- --- --- 0.0% 2.1% 0.0% 0.0% . --- --- --- --- --- --- --- ---

Br -0.56 0.29 0.64 0.79 0.99 1.00 --- --- --- --- --- --- --- 0.0% 2.6% 0.0% 0.0% 0.0% . --- --- --- --- --- --- ---

N(V) 0.11 -0.25 -0.04 -0.12 -0.03 0.00 1.00 --- --- --- --- --- --- 42.3% 5.4% 77.1% 35.0% 83.0% 97.6% . --- --- --- --- --- ---

S(VI) 0.80 -0.38 -0.04 -0.59 -0.66 -0.66 0.07 1.00 --- --- --- --- --- 0.0% 0.3% 75.6% 0.0% 0.0% 0.0% 57.2% . --- --- --- --- ---

As -0.46 0.27 0.67 0.78 0.89 0.88 0.07 -0.48 1.00 --- --- --- --- 0.0% 4.1% 0.0% 0.0% 0.0% 0.0% 61.3% 0.0% . --- --- --- ---

As(III) -0.46 0.32 0.67 0.78 0.90 0.88 0.06 -0.51 0.99 1.00 --- --- --- 0.0% 1.1% 0.0% 0.0% 0.0% 0.0% 62.5% 0.0% 0.0% . --- --- ---

As(V) -0.49 0.22 0.55 0.77 0.82 0.82 0.00 -0.53 0.91 0.87 1.00 --- --- 0.0% 9.3% 0.0% 0.0% 0.0% 0.0% 98.4% 0.0% 0.0% 0.0% . --- ---

MMAA -0.57 0.21 0.31 0.67 0.77 0.77 0.05 -0.69 0.72 0.70 0.86 1.00 --- 0.0% 10.5% 1.5% 0.0% 0.0% 0.0% 71.4% 0.0% 0.0% 0.0% 0.0% . ---

DMAA -0.35 -0.04 0.58 0.56 0.67 0.67 0.05 -0.22 0.77 0.71 0.68 0.45 1.00 0.6% 75.1% 0.0% 0.0% 0.0% 0.0% 70.9% 8.5% 0.0% 0.0% 0.0% 0.0% .

Al 0.84 -0.45 0.06 -0.57 -0.53 -0.53 0.03 0.86 -0.41 -0.42 -0.44 -0.59 -0.17 0.0% 0.0% 64.0% 0.0% 0.0% 0.0% 82.9% 0.0% 0.1% 0.1% 0.0% 0.0% 19.8%B -0.56 0.33 0.67 0.80 0.96 0.95 -0.15 -0.66 0.82 0.82 0.76 0.72 0.65 0.0% 1.1% 0.0% 0.0% 0.0% 0.0% 24.4% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

Ba 0.41 -0.30 0.39 -0.11 0.16 0.17 0.12 0.15 0.17 0.19 0.13 0.09 0.21 0.1% 2.1% 0.2% 40.6% 22.3% 18.6% 36.3% 25.1% 19.1% 15.5% 31.7% 51.6% 9.9%

Be 0.41 -0.24 0.20 -0.26 -0.05 -0.07 0.03 0.35 -0.12 -0.12 -0.18 -0.22 0.11 0.1% 6.2% 13.3% 4.4% 72.1% 57.8% 83.9% 0.7% 34.9% 35.5% 16.9% 8.7% 42.2%

Ca -0.10 -0.22 0.30 -0.02 0.28 0.27 0.13 -0.07 0.10 0.09 -0.06 0.03 0.24 44.2% 9.1% 2.0% 89.0% 2.8% 3.4% 31.9% 59.1% 44.2% 50.8% 64.7% 81.3% 6.0%

Cd 0.34 -0.05 0.32 -0.10 -0.09 -0.11 -0.12 0.33 -0.02 -0.04 -0.08 -0.29 0.03 0.7% 68.0% 1.2% 43.1% 50.6% 40.1% 37.7% 1.1% 90.4% 77.8% 52.8% 2.6% 82.4%

Co 0.26 -0.04 0.40 0.01 0.16 0.15 0.30 0.04 0.10 0.15 -0.02 -0.07 -0.04 4.8% 74.3% 0.2% 91.4% 21.6% 24.1% 2.1% 78.4% 43.3% 24.6% 88.2% 59.8% 76.9%

Cr 0.47 -0.04 0.36 -0.11 -0.14 -0.14 0.31 0.45 -0.04 -0.03 -0.13 -0.25 -0.01 0.0% 77.1% 0.5% 40.3% 28.6% 30.4% 1.4% 0.0% 73.8% 82.7% 30.6% 5.8% 96.4%

Cu 0.15 -0.06 -0.03 -0.08 -0.05 -0.07 0.16 0.03 -0.01 -0.01 -0.04 0.03 -0.02 26.1% 67.0% 83.9% 52.4% 69.1% 58.0% 22.0% 83.7% 94.1% 91.4% 77.9% 79.6% 86.5%

K -0.17 -0.09 0.61 0.36 0.68 0.66 -0.01 -0.23 0.54 0.55 0.41 0.34 0.59 18.6% 47.8% 0.0% 0.5% 0.0% 0.0% 96.9% 8.2% 0.0% 0.0% 0.1% 0.8% 0.0%

Li -0.62 0.37 0.60 0.84 0.94 0.93 -0.09 -0.59 0.88 0.88 0.81 0.74 0.63 0.0% 0.4% 0.0% 0.0% 0.0% 0.0% 49.7% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

Mg 0.56 -0.52 -0.21 -0.70 -0.54 -0.54 0.18 0.59 -0.52 -0.54 -0.57 -0.56 -0.15 0.0% 0.0% 11.1% 0.0% 0.0% 0.0% 17.4% 0.0% 0.0% 0.0% 0.0% 0.0% 24.6%

224 Appendix

{H+} temp cond F Cl Br N(V) S(VI) As As(III) As(V) MMAA DMAA

Mn 0.51 -0.43 -0.06 -0.58 -0.38 -0.39 0.23 0.55 -0.38 -0.40 -0.48 -0.49 -0.05 0.0% 0.1% 66.0% 0.0% 0.2% 0.2% 8.2% 0.0% 0.2% 0.2% 0.0% 0.0% 68.0%

Mo -0.55 0.43 0.44 0.74 0.73 0.71 -0.17 -0.64 0.58 0.59 0.60 0.59 0.34 0.0% 0.1% 0.1% 0.0% 0.0% 0.0% 19.8% 0.0% 0.0% 0.0% 0.0% 0.0% 0.8%

Na -0.71 0.38 0.57 0.89 0.95 0.93 -0.08 -0.66 0.88 0.87 0.83 0.76 0.66 0.0% 0.3% 0.0% 0.0% 0.0% 0.0% 56.6% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%

Ni 0.59 -0.29 0.27 -0.32 -0.21 -0.23 0.24 0.56 -0.05 -0.10 -0.11 -0.23 0.15 0.0% 2.4% 3.9% 1.2% 10.5% 7.3% 6.4% 0.0% 72.3% 45.2% 41.0% 8.1% 25.5%

Pb 0.19 0.20 0.28 0.05 0.04 0.04 -0.02 0.17 0.08 0.11 -0.08 -0.19 -0.06 14.8% 13.1% 3.2% 69.7% 75.4% 76.5% 88.5% 20.1% 52.2% 41.8% 56.5% 15.0% 66.5%

Sr 0.18 -0.38 0.21 -0.21 0.02 0.02 0.13 0.10 -0.11 -0.12 -0.24 -0.16 0.16 18.1% 0.3% 10.3% 11.2% 87.7% 87.5% 30.7% 43.7% 40.1% 35.8% 6.9% 21.9% 23.2%V 0.44 0.02 0.27 -0.23 -0.22 -0.25 0.16 0.47 -0.21 -0.19 -0.31 -0.36 -0.21 0.0% 89.3% 3.9% 7.7% 8.8% 5.8% 23.3% 0.0% 10.2% 14.2% 1.5% 0.4% 11.1%

Zn 0.52 -0.26 0.19 -0.34 -0.12 -0.14 0.36 0.38 -0.04 -0.03 -0.10 -0.15 0.08 0.0% 4.7% 14.6% 0.8% 37.0% 28.8% 0.5% 0.3% 78.6% 79.5% 45.1% 25.4% 54.0%

SiO2 -0.18 0.25 0.65 0.52 0.64 0.60 -0.29 -0.32 0.48 0.51 0.43 0.37 0.34 17.5% 5.7% 0.0% 0.0% 0.0% 0.0% 2.2% 1.2% 0.0% 0.0% 0.1% 0.4% 0.9%

S(-II) -0.04 -0.02 -0.30 -0.15 -0.30 -0.30 0.16 0.14 -0.18 -0.18 -0.24 -0.28 0.03 75.0% 88.2% 2.0% 26.3% 1.9% 1.8% 21.3% 28.8% 17.0% 16.0% 6.2% 3.2% 79.6%

Fe(II) 0.67 -0.34 -0.02 -0.64 -0.43 -0.42 0.18 0.62 -0.33 -0.31 -0.41 -0.49 -0.10 0.0% 0.9% 85.8% 0.0% 0.1% 0.1% 15.9% 0.0% 1.0% 1.4% 0.1% 0.0% 43.9%

Fe(III) 0.63 -0.38 0.22 -0.40 -0.21 -0.21 0.23 0.51 -0.10 -0.11 -0.11 -0.23 0.06 0.0% 0.3% 9.3% 0.1% 11.1% 10.3% 7.8% 0.0% 45.2% 38.2% 41.2% 8.1% 62.6%

TIC 0.02 -0.23 -0.24 -0.18 -0.36 -0.36 -0.01 0.12 -0.22 -0.25 -0.24 -0.26 0.03 88.9% 7.2% 6.0% 16.9% 0.5% 0.5% 93.5% 34.2% 9.8% 5.3% 6.4% 4.9% 82.6%

TOC 0.57 -0.31 -0.29 -0.65 -0.60 -0.58 0.41 0.60 -0.48 -0.48 -0.53 -0.51 -0.30 0.0% 1.7% 2.3% 0.0% 0.0% 0.0% 0.1% 0.0% 0.0% 0.0% 0.0% 0.0% 2.0%

Al B Ba Be Ca Cd Co Cr Cu K Li Mg Mn Al 1.00 --- --- --- --- --- --- --- --- --- --- --- --- . --- --- --- --- --- --- --- --- --- --- --- --- B -0.50 1.00 --- --- --- --- --- --- --- --- --- --- --- 0.0% . --- --- --- --- --- --- --- --- --- --- ---

Ba 0.39 0.19 1.00 --- --- --- --- --- --- --- --- --- --- 0.2% 15.2% . --- --- --- --- --- --- --- --- --- ---

Be 0.54 0.02 0.45 1.00 --- --- --- --- --- --- --- --- --- 0.0% 90.1% 0.0% . --- --- --- --- --- --- --- --- ---

Ca -0.03 0.28 0.12 0.22 1.00 --- --- --- --- --- --- --- --- 84.1% 3.2% 37.7% 8.8% . --- --- --- --- --- --- --- ---

Cd 0.32 -0.01 0.15 0.39 -0.02 1.00 --- --- --- --- --- --- --- 1.2% 91.2% 24.4% 0.2% 86.7% . --- --- --- --- --- --- ---

Co 0.25 0.17 0.35 0.40 0.21 0.36 1.00 --- --- --- --- --- --- 5.6% 19.0% 0.6% 0.1% 10.0% 0.4% . --- --- --- --- --- ---

Cr 0.46 -0.14 0.30 0.25 0.22 0.21 0.43 1.00 --- --- --- --- --- 0.0% 27.5% 1.9% 5.2% 9.3% 10.9% 0.1% . --- --- --- --- ---

Cu 0.15 -0.04 0.18 0.32 0.04 0.02 0.21 0.07 1.00 --- --- --- --- 24.7% 77.3% 17.6% 1.3% 78.9% 89.5% 10.0% 61.1% . --- --- --- ---

K -0.04 0.69 0.47 0.51 0.53 0.17 0.33 0.08 0.06 1.00 --- --- --- 74.6% 0.0% 0.0% 0.0% 0.0% 19.8% 1.1% 56.8% 62.3% . --- --- ---

Li -0.55 0.89 -0.01 -0.15 0.23 -0.14 0.07 -0.16 -0.11 0.58 1.00 --- --- 0.0% 0.0% 96.1% 23.8% 8.3% 27.2% 61.4% 23.7% 40.4% 0.0% . --- ---

Appendix 225

Al B Ba Be Ca Cd Co Cr Cu K Li Mg Mn Mg 0.63 -0.52 0.38 0.56 0.39 0.16 0.12 0.34 0.23 0.08 -0.61 1.00 ---

0.0% 0.0% 0.3% 0.0% 0.2% 22.9% 36.5% 0.8% 8.3% 53.8% 0.0% . --- Mn 0.61 -0.36 0.42 0.68 0.48 0.24 0.29 0.46 0.30 0.25 -0.46 0.94 1.00

0.0% 0.4% 0.1% 0.0% 0.0% 6.4% 2.3% 0.0% 1.9% 5.3% 0.0% 0.0% . Mo -0.58 0.73 -0.13 -0.13 0.05 0.03 0.16 -0.28 -0.05 0.30 0.72 -0.63 -0.56

0.0% 0.0% 30.7% 31.8% 68.0% 84.8% 23.0% 3.0% 68.8% 2.0% 0.0% 0.0% 0.0% Na -0.62 0.91 -0.06 -0.22 0.20 -0.14 0.01 -0.17 -0.15 0.54 0.96 -0.63 -0.50 0.0% 0.0% 65.7% 9.4% 11.7% 27.3% 93.1% 19.0% 24.2% 0.0% 0.0% 0.0% 0.0%

Ni 0.61 -0.22 0.31 0.47 0.28 0.31 0.20 0.53 0.29 0.10 -0.26 0.56 0.62 0.0% 9.5% 1.7% 0.0% 3.3% 1.7% 12.8% 0.0% 2.2% 46.5% 4.6% 0.0% 0.0%

Pb 0.12 0.05 0.04 0.09 0.08 0.46 0.44 0.29 -0.13 0.09 0.03 -0.09 0.03 34.6% 70.5% 73.7% 49.0% 56.1% 0.0% 0.0% 2.5% 30.7% 50.6% 83.0% 48.8% 81.3%

Sr 0.25 0.05 0.46 0.32 0.78 0.02 0.26 0.37 0.13 0.44 -0.11 0.65 0.67 5.4% 71.6% 0.0% 1.3% 0.0% 88.3% 4.5% 0.3% 30.9% 0.0% 41.8% 0.0% 0.0% V 0.44 -0.21 0.17 0.27 0.30 0.30 0.41 0.81 -0.05 0.01 -0.24 0.33 0.45 0.0% 10.1% 20.1% 3.6% 1.9% 2.1% 0.1% 0.0% 72.5% 94.6% 6.8% 1.0% 0.0%

Zn 0.53 -0.12 0.53 0.76 0.17 0.32 0.50 0.49 0.46 0.33 -0.21 0.61 0.71 0.0% 35.1% 0.0% 0.0% 18.2% 1.3% 0.0% 0.0% 0.0% 0.9% 11.0% 0.0% 0.0%

SiO2 -0.15 0.72 0.18 0.37 0.17 0.30 0.30 -0.05 0.14 0.60 0.56 -0.30 -0.17 24.2% 0.0% 17.4% 0.4% 20.4% 2.1% 2.2% 69.5% 28.2% 0.0% 0.0% 1.9% 20.0%

S(-II) 0.07 -0.32 -0.26 -0.12 0.02 -0.19 -0.32 0.14 -0.09 -0.27 -0.24 0.14 0.10 59.3% 1.1% 4.6% 35.3% 86.1% 14.0% 1.4% 30.2% 48.2% 4.0% 6.0% 29.0% 44.5%

Fe(II) 0.72 -0.41 0.52 0.67 0.16 0.26 0.24 0.45 0.30 0.12 -0.49 0.78 0.83 0.0% 0.1% 0.0% 0.0% 21.7% 4.4% 6.5% 0.0% 2.2% 36.4% 0.0% 0.0% 0.0%

Fe(III) 0.61 -0.19 0.64 0.66 0.16 0.44 0.41 0.42 0.34 0.24 -0.32 0.62 0.68 0.0% 15.2% 0.0% 0.0% 21.8% 0.0% 0.1% 0.1% 0.8% 6.0% 1.4% 0.0% 0.0%

TIC 0.01 -0.36 -0.32 -0.28 -0.03 -0.09 -0.43 0.00 -0.15 -0.36 -0.30 0.10 -0.05 94.4% 0.5% 1.3% 3.0% 83.2% 48.6% 0.1% 98.5% 24.3% 0.5% 1.9% 46.6% 72.9%

TOC 0.50 -0.65 0.23 0.18 0.11 0.08 0.04 0.38 0.06 -0.26 -0.61 0.66 0.58 0.0% 0.0% 8.1% 16.1% 39.4% 54.9% 79.0% 0.3% 63.6% 4.2% 0.0% 0.0% 0.0%

Mo Na Ni Pb Sr V Zn SiO2 S(-II) Fe(II) Fe(III) TIC TOC

Mo 1.00 --- --- --- --- --- --- --- --- --- --- --- --- . --- --- --- --- --- --- --- --- --- --- --- ---

Na 0.72 1.00 --- --- --- --- --- --- --- --- --- --- --- 0.0% . --- --- --- --- --- --- --- --- --- --- ---

Ni -0.32 -0.31 1.00 --- --- --- --- --- --- --- --- --- --- 1.4% 1.6% . --- --- --- --- --- --- --- --- --- ---

Pb 0.04 0.02 0.05 1.00 --- --- --- --- --- --- --- --- --- 78.1% 87.4% 70.7% . --- --- --- --- --- --- --- --- ---

Sr -0.20 -0.10 0.39 -0.03 1.00 --- --- --- --- --- --- --- --- 12.1% 45.4% 0.2% 83.9% . --- --- --- --- --- --- --- --- V -0.26 -0.24 0.50 0.40 0.38 1.00 --- --- --- --- --- --- --- 4.1% 6.1% 0.0% 0.2% 0.3% . --- --- --- --- --- --- ---

Zn -0.27 -0.29 0.69 0.07 0.34 0.42 1.00 --- --- --- --- --- --- 3.7% 2.5% 0.0% 60.4% 0.7% 0.1% . --- --- --- --- --- ---

SiO2 0.68 0.53 -0.02 0.09 0.00 -0.05 0.11 1.00 --- --- --- --- --- 0.0% 0.0% 87.4% 51.6% 97.7% 71.6% 39.3% . --- --- --- --- ---

S(-II) -0.37 -0.18 0.21 -0.02 0.05 0.18 -0.04 -0.50 1.00 --- --- --- --- 0.3% 16.5% 11.5% 87.3% 70.4% 16.2% 78.9% 0.0% . --- --- --- ---

Fe(II) -0.57 -0.57 0.61 0.08 0.39 0.42 0.77 -0.12 0.10 1.00 --- --- --- 0.0% 0.0% 0.0% 54.5% 0.2% 0.1% 0.0% 37.4% 44.2% . --- --- ---

226 Appendix

Mo Na Ni Pb Sr V Zn SiO2 S(-II) Fe(II) Fe(III) TIC TOCFe(III) -0.29 -0.39 0.65 0.14 0.36 0.31 0.81 0.08 -0.18 0.77 1.00 --- ---

2.7% 0.2% 0.0% 27.5% 0.4% 1.4% 0.0% 54.1% 16.2% 0.0% . --- --- TIC -0.36 -0.21 0.02 -0.08 0.02 0.03 -0.24 -0.45 0.68 -0.10 -0.24 1.00 ---

0.5% 10.5% 88.4% 53.8% 85.9% 84.5% 6.9% 0.0% 0.0% 43.4% 6.9% . --- TOC -0.64 -0.63 0.48 0.15 0.31 0.35 0.37 -0.55 0.30 0.60 0.44 0.15 1.00

0.0% 0.0% 0.0% 25.6% 1.6% 0.6% 0.3% 0.0% 1.8% 0.0% 0.0% 26.3% .

Appendix 15 Saturation indices for albite, potassium feldspar (adularia), muscovite (Kmica), Mg-clinochlore

(chlorite 14A), and silica (quartz, and SiO2(a)); these minerals determine the partial or full equilibrium of waters

according to the simplified ternary diagram after Giggenbach (1988); site names in black are in full equilibrium

according to the ternary diagrams in Figure 46, Figure 49, and Figure 60 , site names in italic are in partial equi-

librium

Site name Albite Adularia Kmica Chlorite14A Quartz SiO2(a)

YNPGG02-2W -0.20 0.05 7.36 -20.07 0.84 -0.13 YNPGG04-1W -4.38 -3.92 0.20 -42.24 0.71 -0.23 YNPGG04-2W -5.49 -5.13 -3.27 -44.45 0.72 -0.22 YNPGG04-4W -5.35 -4.98 -2.37 -43.31 0.66 -0.27 YNPGG05-2W -9.18 -7.56 -7.82 -60.07 0.79 -0.23 YNPGG05-4W -9.54 -7.56 -8.30 -61.19 0.81 -0.25 YNPGG06-2W 0.69 1.28 7.33 -9.60 0.89 -0.04 YNPGG07-2W -0.43 0.16 6.29 -18.28 0.79 -0.14 YNPGG07-4W -0.70 -0.12 5.63 -19.16 0.78 -0.15 YNPGG08-4W -8.97 -8.27 -11.22 -62.44 0.90 -0.10 YNPHL01-1W -7.33 -5.40 -6.87 -59.41 1.40 0.09 YNPHL01-2W -6.37 -4.70 -5.65 -55.85 1.48 0.22 YNPHL01-3W -6.92 -5.10 -6.94 -57.91 1.45 0.18 YNPHL02-1W 0.92 2.72 12.09 -17.07 1.28 0.04 YNPHL03-3W -6.70 -4.81 -4.50 -51.15 1.22 0.13 YNPLG01-2W 0.07 0.32 5.30 -6.81 0.78 -0.14 YNPLG02-2W -0.25 -0.07 2.49 3.91 0.61 -0.30 YNPLG03-2W -0.09 0.18 5.01 -6.95 0.75 -0.16 YNPNL01-1W -9.77 -7.74 -12.40 -65.90 1.34 0.04 YNPNL02-1W -8.31 -7.09 -8.03 -58.33 0.97 -0.08 YNPNL02-2W -7.98 -6.94 -7.48 -56.18 0.89 -0.10 YNPNL02-3W -12.17 -10.12 -12.96 -65.65 0.49 -0.58 YNPNL03-1W -8.11 -6.49 -4.74 -45.54 0.56 -0.43 YNPNL03-2W -7.99 -6.52 -4.77 -46.31 0.62 -0.33 YNPNL03-3W -8.82 -7.25 -6.36 -50.81 0.64 -0.32 YNPNL04-1W -8.89 -6.88 -8.00 -54.57 0.94 -0.23 YNPNL04-2W -10.22 -8.56 -11.46 -59.03 0.91 -0.16 YNPNL04-3W -9.53 -7.84 -8.90 -57.18 0.81 -0.22 YNPNL05-1W -8.59 -7.04 -8.84 -60.17 1.10 0.09 YNPNL05-2W -8.28 -6.72 -7.78 -57.88 1.06 0.11 YNPRH01-1W -6.75 -5.44 -5.72 -53.47 1.13 0.09 YNPRH01-2W -6.43 -5.19 -4.68 -51.57 1.06 0.04 YNPRH01-3W -6.70 -5.43 -4.98 -51.96 1.04 0.03 YNPRH03-1W -8.47 -6.91 -10.92 -66.80 1.48 0.30 YNPRH04-1W -8.13 -6.64 -9.75 -64.39 1.40 0.25

Appendix 227

Site name Albite Adularia Kmica Chlorite14A Quartz SiO2(a)

YNPRH04-2W -8.11 -6.64 -9.87 -64.72 1.42 0.27 YNPRH04-3W -8.31 -6.87 -10.10 -64.65 1.38 0.26 YNPRH05-1W -7.43 -6.49 -6.91 -54.35 0.79 -0.25 YNPRH05-3W -8.38 -7.48 -8.95 -49.00 0.56 -0.37 YNPRH07-2W -10.67 -9.06 -9.62 -61.65 0.40 -0.55 YNPRH08-1W -8.29 -7.48 -13.13 -59.98 1.21 0.20 YNPRH08-4W -0.18 0.50 6.31 -22.66 1.10 0.15 YNPRH09-1W 0.67 1.56 7.37 -6.61 0.97 -0.01 YNPRH10-4W -5.32 -4.47 -4.06 -39.83 0.95 0.03 YNPRH11-3W -8.85 -7.76 -9.83 -59.06 0.77 -0.27 YNPRH13-3W -6.22 -5.39 -8.39 -48.49 1.11 0.13 YNPRH14-3W -5.35 -4.72 -6.47 -41.59 1.00 0.05 YNPRH15-3W -9.72 -8.47 -15.73 -68.60 1.35 0.16 YNPRH16-4W -9.30 -8.05 -15.00 -67.36 1.34 0.16 YNPRH17-4W -7.28 -6.21 -8.26 -54.56 1.10 0.09 YNPRH18-4W -6.64 -5.43 -3.56 -47.13 0.82 -0.11 YNPRH19-4W -6.82 -6.01 -8.52 -47.02 1.02 0.10 YNPRH20-4W -7.45 -6.41 -11.01 -53.91 1.30 0.28 YNPRH21-4W -8.81 -7.90 -13.28 -58.61 1.12 0.14 YNPRH22-4W -7.12 -5.89 -5.88 -55.78 1.07 -0.01 YNPRH24-4W -8.16 -7.03 -11.46 -61.66 1.31 0.21 YNPRH25-4W -6.83 -5.92 -7.10 -46.34 0.91 -0.01 YNPRH26-4W -6.06 -5.01 -4.24 -50.43 0.99 0.00 YNPRH27-4W -6.52 -6.14 -8.48 -42.33 0.84 -0.09 YNPRH28-4W -7.55 -6.56 -10.62 -47.69 1.03 0.04

Appendix 16 Reservoir temperatures calculated from the following geothermometers: Si (Fournier and Potter

1982), K-Mg (Giggenbach 1988), Na-K (Arnórsson et al. 1998), Na-K-Ca (Fournier and Truesdell 1973)*

Analysis-code-water Cluster No. Temp [°C] from Si geotherm.

Temp [°] from K-Mg geotherm.

Temp [°C] from Na-K geotherm.

Temp [°C] from Na-K-Ca geotherm.

YNPGG04-2W 5 165 136 140 228 YNPLG01-2W 8 181 168 128 235 YNPLG02-2W 8 175 163 119 226 YNPLG03-2W 8 177 168 131 238 YNPGG02-2W 11 171 148 117 210 YNPGG06-2W 11 193 215 178 273 YNPGG07-2W 11 178 211 178 271 YNPGG07-4W 11 175 209 175 266 YNPRH27-4W 13 183 154 143 239 YNPRH08-1W 14 212 199 191 281 YNPRH08-4W 14 216 215 187 276 YNPRH13-3W 14 207 226 206 294 YNPRH14-3W 14 200 211 177 270

* Si-geothermometer: T = -53.5 + 0.11236 ⋅ S – 0.5559 ⋅ 10-4 ⋅ S2 + 0.1772 ⋅ 10-7 ⋅ S3 + 88.390 log S (S = SiO2 [mg/L]) K-Mg-geotherm.: T = 4410 : (14.0 - log X) - 273.15 (X = log (K[mg/L]2 / Mg[mg/L]) Na-K-geotherm.: T = 733.6 - 770.551 ⋅ Y + 378.189 ⋅ Y2 - 95.753 ⋅ Y3 + 9.544 ⋅ Y4

(Y = log Na [mol/L] / K [mol/L]) Na-K-Ca-geotherm.: T = 1647 : (log (Na/K) + 1/3 ⋅ log (Ca0.5/Na) + 2.24) – 273.15 (concentrations in mol/L)

228 Appendix

Appendix 17 Mann-Whitney Test for the difference between the two major groups for 60 water samples derived

from cluster analysis; significance levels > 1% (not accepted as significant) are marked in bold

Mann-Whitney U Wilcoxon W Z significance level (2-tailed) {H+} 130.5 508.5 -4.7 0.00%

temperature 220 781 -3.4 0.08% conductivity 193.5 754.5 -3.7 0.02%

F 36 597 -6.1 0.00% Cl 0 561 -6.6 0.00% Br 0 561 -6.6 0.00%

N(V) 433 811 -0.2 82.00% S(VI) 59 437 -5.7 0.00%

As(III) 22 583 -6.3 0.00% As(V) 60 621 -5.7 0.00%

MMAA 64.5 625.5 -5.7 0.00% DMAA 178.5 739.5 -4.0 0.01%

Al 115.5 493.5 -4.9 0.00% B 19 580 -6.3 0.00% Ba 382 943 -0.9 35.00% Be 324 702 -1.9 5.40% Ca 376 937 -1.0 30.00% Cd 378 756 -1.8 7.20% Co 392.5 953.5 -0.8 41.00% Cr 421 799 -0.6 54.00% Cu 412 790 -0.6 53.00% K 209 770 -3.5 0.04% Li 5 566 -6.5 0.00% Mg 143 521 -4.5 0.00% Mn 194 572 -3.7 0.02% Mo 121 682 -5.2 0.00% Na 1 562 -6.6 0.00% Ni 293 671 -2.5 1.30% Pb 430.5 991.5 -0.4 71.00% Sr 439 817 -0.1 92.00% V 406 784 -1.2 22.00% Zn 356 734 -1.4 17.00%

SiO2 226 787 -3.3 0.11% S(-II) 375 753 -1.1 29.00% Fe(II) 197.5 575.5 -3.7 0.02% Fe(III) 281.5 659.5 -2.4 1.50%

TIC 364 742 -1.2 23.00% TOC 173.5 551.5 -4.0 0.01%

Appendix 229

Appendix 18 Dendrogram of cluster analysis for 38 samples with combined data from gas analysis (As (from HG-AAS), Al, B, Ba, Cu, Fe, K, Li, Si, Sr, Zn, S) and water analysis (pH, temperature, conductivity, F, Cl, Br, N(V), S(VI), As, As(III), As(V), MMAA, DMAA, Al, B, Ba, Be, Ca, Cd, Co, Cr, Cu, K, Li, Mg, Mn, Mo, Na, Ni, Pb, Sr, V, Zn, SiO2, S(-II), Fe(II), Fe(III), TIC, TOC) (all variables standardized to a range of 0 to 1 and rescaled before cluster analysis; Ward´s method, program: SPSS for Windows 11.0)*

Rescaled distance 0 5 10 15 20 25 +---------+---------+---------+---------+---------+ YNPGG05-4G-NaOCl-wGH òø YNPNL02-3G-NaOCl-wGH òôòø YNPNL04-1G-NaOCl-wGH ò÷ ùòø YNPHL02-1G-NaOCl-wGH òûò÷ ùòòòø YNPNL01-1G-NaOCl-wGH ò÷ ó ó YNPNL04-3G-NaOCl-wGH òòòòò÷ ó YNPGG05-2G-NaOCl-wGH òø ó YNPNL02-2G-NaOCl-wGH òôòø ó YNPNL03-2G-NaOCl-wGH ò÷ ùòòòòòôòø YNPNL04-2G-NaOCl-wGH òòò÷ ó ó YNPGG04-2G-NaOCl-wGH òûòø ó ùòòòòòòòòòòòòòø YNPGG04-4G-NaOCl-wGH ò÷ ùòòòòò÷ ó ó YNPRH01-3G-NaOCl-wGH òòò÷ ó ó YNPNL03-1G-NaOCl-wGH òòòòòòòòòòò÷ ó YNPRH01-2G-NaOCl-wGH òûòòòø ùòòòòòòòòòòòòòòòòòø YNPRH04-2G-NaOCl-wGH ò÷ ùòø ó ó YNPHL01-2G-NaOCl-wGH òòòòò÷ ùòòòòòòòø ó ó YNPRH05-1G-NaOCl-wGH òòòòòòò÷ ó ó ó YNPHL01-1G-NaOCl-wGH òòòòòûòòòòòø ùòòòòòòòòò÷ ó YNPHL01-3G-NaOCl-wGH òòòòò÷ ó ó ó YNPRH03-1G-NaOCl-wGH òø ùòòò÷ ó YNPRH04-1G-NaOCl-wGH òôòòòòòø ó ùòòòòòø YNPRH04-3G-NaOCl-wGH ò÷ ùòòò÷ ó ó YNPRH01-1G-NaOCl-wGH òòòòòòò÷ ó ó YNPLG01-2G-NaOCl-wGH òûòòòø ó ó YNPLG02-2G-NaOCl-wGH ò÷ ùòòòòòòòòòø ó ó YNPLG03-2G-NaOCl-wGH òòòòò÷ ó ó ó YNPGG02-2G-NaOCl-wGH òûòø ùòòòòòòòòòòòòòòòòòòòòòø ó ó YNPGG07-4G-NaOCl-wGH ò÷ ùòòòø ó ó ó ó YNPRH09-1G-NaOCl-wGH òòò÷ ùòòòòòòò÷ ùòòòòò÷ ó YNPRH08-1G-NaOCl-wGH òòòòòòò÷ ó ó YNPRH05-3G-NaOCl-wGH òòòòòòòòòòòûòòòòòòòòòòòòòø ó ó YNPRH11-4G-NaOCl-wGH òòòòòòòòòòò÷ ùòòòòòòòòòòò÷ ó YNPRH20-4G-NaOCl-wGH òòòòòòòûòòòòòø ó ó YNPRH22-4G-NaOCl-wGH òòòòòòò÷ ùòòòòòòòòòòò÷ ó YNPRH07-2G-NaOCl-wGH òòòòòòòòòòòòò÷ ó YNPNL05-1G-NaOCl-wGH òûòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòòò÷ YNPNL05-2G-NaOCl-wGH ò÷

* for gas samples where no contemporaneous water samples existed the following cases were assigned to the respective subgroups: A: YNPNL06-1G; B: YNPGG03-2G, 4G; YNPRH01-3G setup 2; YNPRH02-1G; D: YNPRH07-1G; YNPRH12-4G; YNPRH22-4G setup 2; YNPRH23-4G; F: YNPRH05-4G setup 1 and 2; YNPRH06-1G, 2G; E: YNPGG01-2G; YNPGG02-4G; YNPLG01-1G; YNPRH08-2G

group B

group C

group E

group F

group D

group A

230 Appendix

Appendix 19 Kruskall-Wallis test for overall difference between the 6 subgroups derived from cluster analysis of 38 samples with both gas and water data after Ward´s method; df = degree of freedom = 5; significance levels > 5% (not accepted as significant) are marked in bold

Para

met

ers f

rom

ga

s sam

ples

Chi

-Squ

are

sign

ifica

nce

leve

l

Para

met

ers f

rom

gas

an

d w

ater

sam

ples

Chi

-Squ

are

sign

ifica

nce

leve

l

Para

met

ers f

rom

w

ater

sam

ples

Chi

-Squ

are

sign

ifica

nce

leve

l

As (AAS) nGH* 15 0.9% Zn wGH 5 46.9% Cd 21 0.1% As (AAS) wGH* 15 0.9% S(VI) nGH 3 76.4% Co 11 4.3%

Al nGH 17 0.4% S(VI) wGH 3 74.6% Cr 24 0.0% Al wGH 17 0.4% {H+} 19 0.2% Cu 2 80.9% B nGH 11 5.2% temperature 18 0.3% K 20 0.1% B wGH 11 5.2% conductivity 22 0.1% Li 23 0.0%

Ba nGH 8 13.6% F 21 0.1% Mg 18 0.3%

Ba wGH 9 11.8% Cl 27 0.0% Mn 21 0.1%

Cu nGH 6 30.6% Br 27 0.0% Mo 24 0.0% Cu wGH 7 24.7% N(V) 9 9.6% Na 26 0.0% Fe nGH 12 3.4% S(VI) 21 0.1% Ni 20 0.1% Fe wGH 12 3.4% As 26 0.0% Pb 20 0.1% K nGH 11 5.7% As(III) 25 0.0% Sr 8 18.1% K wGH 11 5.7% As(V) 25 0.0% V 21 0.1% Li nGH 14 1.4% MMAA 17 0.4% Zn 21 0.1% Li wGH 18 0.3% DMAA 26 0.0% SiO2 16 0.6%

SiO2 nGH 8 16.8% Al 23 0.0% S(-II) 6 32.1% SiO2 WGH 8 16.8% B 27 0.0% Fe(II) 26 0.0%

Sr nGH 7 23.3% Ba 29 0.0% Fe(III) 27 0.0% Sr wGH 8 14.2% Be 24 0.0% TIC 8 16.7% Zn nGH 4 50.2% Ca 4 51.9% TOC 19 0.2%

* all parameters for setups with gas hood (setup 1 or 2 in Figure 42); “nGH” behind an element indicates that the concentration in air was calculated from the first end-member model “all As from ambient air” (as-suming a “no gas hood” case); “wGH” indicates the second end-member model “all As from underneath the gas hood” (further explanation see section 4.2.3.2.)

Appendix 231

Appendix 20 Kruskall-Wallis test for overall difference between the 6 subgroups derived from cluster analysis considering only the data of 55 gas samples (no water data) after Ward´s method; df = degree of freedom = 5; significance levels > 5% (not accepted as significant) are marked in bold

Pa

ram

eter

s fro

m

gas s

ampl

es

Chi

-Squ

are

sign

ifica

nce

leve

l

Para

met

ers f

rom

ga

s sam

ples

Chi

-Squ

are

sign

ifica

nce

leve

l

Para

met

ers f

rom

ga

s sam

ples

Chi

-Squ

are

sign

ifica

nce

leve

l

As (AAS) nGH 29 0.00% Cu nGH 6 26.52% SiO2 nGH 9 10.42% As (AAS) wGH 29 0.00% Cu wGH 7 18.83% SiO2 WGH 9 10.42%

Al nGH 21 0.07% Fe nGH 16 0.72% Sr nGH 12 3.85% Al wGH 21 0.07% Fe wGH 16 0.72% Sr wGH 13 2.48% B nGH 20 0.12% K nGH 19 0.16% Zn nGH 7 19.55% B wGH 20 0.12% K wGH 19 0.16% Zn wGH 8 18.04% Ba nGH 19 0.20% Li nGH 22 0.06% S(VI) nGH 3 67.29% Ba wGH 19 0.16% Li wGH 25 0.01% S(VI) wGH 3 66.55%

* all parameters for setups with gas hood (setup 1 or 2 in Figure 42); “nGH” behind an element indicates that the concentration in air was calculated from the first end-member model “all As from ambient air” (assuming a “no gas hood” case); “wGH” stands for calculations from the second end-member model “all As from un-derneath the gas hood” (further explanation see section 4.2.3.2.)

Appendix 21 Assignment of individual water samples to the water and gas subgroups derived by cluster analysis ( Appendix 11; Appendix 18)

Individual water samples

Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2


Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2


Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2

YNPGG01-2G E YNPNL02-2W, 2G 1 1 B YNPRH07-2W, 2G 1 1 D YNPGG02-2W, 2G 2 11 E YNPNL02-3W, 3G 1 1 B YNPRH08-1W, 1G 2 14 E

YNPGG02-4G E YNPNL03-1W, 1G 1 1 B YNPRH08-2G E YNPGG03-2G B YNPNL03-2W, 2G 1 1 B YNPRH08-4W 2 14 YNPGG03-4G B YNPNL03-3W 1 2 YNPRH09-1W, 1G 2 11 E YNPGG04-1W 1 5 YNPNL04-1W, 1G 1 1 B YNPRH10-4W 2 14

YNPGG04-2W, 2G 1 5 B YNPNL04-2W, 2G 1 1 B YNPRH11-3W, 4G 2 10 F YNPGG04-4W, 4G 1 5 B YNPNL04-3W, 3G 1 2 B YNPRH12-4G D

YNPGG05-2W 1 1 B YNPNL05-1W, 1G 1 3 A YNPRH13-3W 2 14 YNPGG05-4W, 4G 1 1 B YNPNL05-2W, 2G 1 3 A YNPRH14-3W 2 14

YNPGG06-2W 2 11 YNPNL06-1G A YNPRH15-3W 2 12 YNPGG07-2W 2 11 YNPRH01-1W, 1G 1 7 C YNPRH16-4W 2 12

YNPGG07-4W, 4G 2 11 E YNPRH01-2W, 2G 1 7 C YNPRH17-4W 2 13 YNPGG08-4W 1 5 YNPRH01-3W, 3G 1 7 B YNPRH18-4W 1 7

YNPHL01-1W, 1G 1 4 C YNPRH02-1G B YNPRH19-4W 2 14 YNPHL01-2W, 2G 1 4 C YNPRH03-1W, 1G 1 6 C YNPRH20-4W, 4G 2 14 D

232 Appendix


Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2


Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2


Wat

er ty

pe

Wat

er su

bgro

up1

Gas

subg

roup

2

YNPHL01-3W, 3G 1 4 C YNPRH04-1W, 1G 1 6 C YNPRH21-4W 2 14 YNPHL02-1W, 1G 1 4 B YNPRH04-2W, 2G 1 6 C YNPRH22-4W, 4G 1 7 D

YNPHL03-3W 1 1 YNPRH04-3W, 3G 1 6 C YNPRH23-4G D YNPLG01-1G E YNPRH05-1W, 1G 2 9 C YNPRH24-4W 2 13

YNPLG01-2W, 2G 2 8 E YNPRH05-3W, 3G 2 9 F YNPRH25-4W 2 13 YNPLG02-2W, 2G 2 8 E YNPRH05-4G F YNPRH26-4W 2 13 YNPLG03-2W, 2G 2 8 E YNPRH06-1G F YNPRH27-4W 2 13 YNPNL01-1W, 1G 1 4 B YNPRH06-2G F YNPRH28-4W 2 13

YNPNL02-1W 1 1 YNPRH07-1G D

1) Water subgroup

Subgroup Subgroup Name No. of cases Type I steam-heated waters

1 Low Cl hot springs and fumaroles at Hazle Lake, Nymph Lake, Gibbon Geyser Basin and Ragged Hills

11

2 Frying Pan Spring during September sampling 2 3 Fumaroles at Nymph Lake from 2003 2 4 Lake samples 5 5 Hot springs at Gibbon Geyser Basin 4 6 “Verde Pool” 4 7 “Milky Way” - “Kaolin Spring” complex 5

Type II deep geothermal waters 8 Lower Geyser Basin alkaline hot springs 3 9 Ragged Hills No. 5 2 10 Ragged Hills No. 11 1 11 Near neutral pH hot springs at Gibbon Geyser Basin 5 12 Warm springs Ragged Hills 2 13 Shallow hot springs Ragged Hills 6 14 Deeper hot springs Ragged Hills 8

2) Gas subgroup

Subgroup Subgroup Name No. of cases A lowest total volatile concentrations (Fumaroles at Nymph Lake from 2003) 3 B low to medium volatile As, high volatile Si, S concentrations due to high water tem-

peratures (low Cl hot springs, fumaroles; hot springs at Gibbon GB, “Verde Pool”) 18

C medium to high volatile As, low volatile Si, S concentrations due to low water tem-peratures (Lakes and “Milky Way” - “Kaolin Spring” complex

10

D highest volatile As concentrations of all samples taken 7 E low volatile As and S concentrations (Lower Geyser Basin alkaline hot springs; near

neutral pH hot springs at Gibbon Geyser Basin and deeper hot springs Ragged Hills) 11

F high volatile As, medium Si and S concentrations (Ragged Hills No. 5 and No. 11) 6

Appendix 233

Appendix 22 Bivariate correlation of volatile metallics considering 55 gas samples (values below detection limit were dismissed, thus for some parameter combination (e.g., Li-Fe) as little as 9 cases were considered; c = Spear-man correlation coefficient, N = number of cases, sig = two tailed level of significance in %)

As Al B Ba Cu Fe K Li SiO2 Sr Zn S(VI) As c 1.00 0.55 0.09 0.41 -0.30 0.46 0.54 0.81 0.27 0.35 -0.34 0.20 N 55 47 18 45 36 31 52 20 29 42 20 43 sig . 0.0% 72.3% 0.6% 7.5% 0.9% 0.0% 0.0% 15.4% 2.1% 14.8% 19.7%

Al c 0.55 1.00 -0.26 0.17 -0.21 0.57 0.59 0.19 0.07 0.14 -0.42 0.09 N 47 47 16 40 34 30 47 18 29 37 14 38 sig 0.0% . 32.7% 30.3% 24.3% 0.1% 0.0% 44.8% 70.9% 39.3% 13.1% 59.5% B c 0.09 -0.26 1.00 0.63 -0.17 -0.25 -0.11 -0.02 0.32 0.20 -0.13 0.39 N 18 16 18 18 15 14 18 11 14 17 14 16 sig 72.3% 32.7% . 0.5% 53.3% 39.2% 65.1% 95.8% 26.0% 44.5% 64.8% 13.4%

Ba c 0.41 0.17 0.63 1.00 -0.33 0.38 0.03 0.05 0.46 -0.13 0.00 0.11 N 45 40 18 45 36 28 45 18 22 42 19 38 sig 0.6% 30.3% 0.5% . 5.3% 4.6% 85.8% 83.6% 3.1% 40.5% 99.4% 51.4%

Cu c -0.30 -0.21 -0.17 -0.33 1.00 -0.33 0.02 -0.76 -0.10 0.42 0.46 0.27 N 36 34 15 36 36 26 36 12 18 36 13 29 sig 7.5% 24.3% 53.3% 5.3% . 10.0% 91.8% 0.5% 68.1% 1.2% 11.7% 15.7%

Fe c 0.46 0.57 -0.25 0.38 -0.33 1.00 0.38 0.57 0.28 -0.02 -0.22 0.39 N 31 30 14 28 26 31 31 9 21 27 10 24 sig 0.9% 0.1% 39.2% 4.6% 10.0% . 3.3% 11.2% 21.8% 91.3% 53.3% 5.9%

K c 0.54 0.59 -0.11 0.03 0.02 0.38 1.00 0.44 0.13 0.45 -0.35 0.19 N 52 47 18 45 36 31 52 20 29 42 19 43 sig 0.0% 0.0% 65.1% 85.8% 91.8% 3.3% . 5.0% 50.8% 0.3% 14.7% 23.2%

Li c 0.81 0.19 -0.02 0.05 -0.76 0.57 0.44 1.00 0.41 0.48 -0.42 0.38 N 20 18 11 18 12 9 20 20 13 16 11 18 sig 0.0% 44.8% 95.8% 83.6% 0.5% 11.2% 5.0% . 16.2% 5.8% 20.1% 11.5%

SiO2 c 0.27 0.07 0.32 0.46 -0.10 0.28 0.13 0.41 1.00 -0.08 0.09 0.19 N 29 29 14 22 18 21 29 13 29 19 10 23 sig 15.4% 70.9% 26.0% 3.1% 68.1% 21.8% 50.8% 16.2% . 73.7% 80.3% 38.1%

Sr c 0.35 0.14 0.20 -0.13 0.42 -0.02 0.45 0.48 -0.08 1.00 -0.35 0.32 N 42 37 17 42 36 27 42 16 19 42 18 35 sig 2.1% 39.3% 44.5% 40.5% 1.2% 91.3% 0.3% 5.8% 73.7% . 15.2% 5.7%

Zn c -0.34 -0.42 -0.13 0.00 0.46 -0.22 -0.35 -0.42 0.09 -0.35 1.00 0.10 N 20 14 14 19 13 10 19 11 10 18 20 18 sig 14.8% 13.1% 64.8% 99.4% 11.7% 53.3% 14.7% 20.1% 80.3% 15.2% . 70.5%

S(VI) c 0.20 0.09 0.39 0.11 0.27 0.39 0.19 0.38 0.19 0.32 0.10 1.00 N 43 38 16 38 29 24 43 18 23 35 18 43 sig 19.7% 59.5% 13.4% 51.4% 15.7% 5.9% 23.2% 11.5% 38.1% 5.7% 70.5% .

234 Appendix

Appendix 23 Bivariate correlation of volatile metallics with water parameters considering 38 samples with gas and water data (values below detection limit were dismissed; c = Spearman correlation coefficient, N = number of cases, sig = two tailed level of significance in %)

A

s (g)

Al (

g)

B (g

)

Ba

(g)

Cu

(g)

Fe (g

)

K (g

)

Li (

g)

SiO

2 (g)

Sr (g

)

Zn

(g)

S(V

I) (g

)

Al c 0.30 0.43 0.56 0.16 -0.01 0.11 0.13 -0.33 -0.14 0.03 -0.27 0.09 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 6.3% 1.9% 7.1% 39.2% 95.5% 67.5% 44.0% 29.7% 57.3% 89.1% 34.2% 66.6%

As c 0.26 0.03 -0.25 0.07 -0.12 0.02 0.07 0.41 0.28 0.04 -0.26 -0.09 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 11.7% 88.4% 46.7% 70.1% 56.5% 94.5% 70.0% 19.1% 25.7% 83.9% 36.6% 66.6%

As(III) c 0.33 0.03 -0.15 0.04 -0.06 0.02 0.08 0.50 0.27 0.12 -0.26 -0.01 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 4.6% 87.3% 66.9% 83.4% 77.6% 92.6% 64.3% 10.1% 28.0% 52.4% 37.9% 95.0%

As(V) c 0.29 0.10 -0.35 0.09 0.05 0.17 0.16 0.40 0.41 0.13 -0.26 0.01 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 7.3% 58.5% 29.8% 61.9% 82.6% 49.9% 36.6% 19.9% 9.4% 48.8% 36.6% 95.9%

MMAA c 0.16 0.08 -0.82 0.16 0.06 0.26 0.18 0.56 0.44 0.23 -0.01 -0.17 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 34.3% 67.7% 0.2% 39.8% 78.6% 28.8% 29.2% 5.8% 6.6% 22.4% 98.2% 38.5%

DMAA c 0.07 0.20 -0.34 0.09 -0.10 -0.15 0.06 -0.06 0.26 -0.09 -0.28 -0.20 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 68.1% 28.5% 31.2% 64.2% 64.9% 55.3% 73.6% 84.6% 30.3% 62.6% 32.6% 31.0%B c 0.09 0.15 -0.65 -0.19 0.09 -0.07 0.18 0.16 0.14 0.14 -0.33 -0.20 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 59.9% 42.1% 2.9% 29.7% 68.9% 77.3% 30.6% 61.8% 58.1% 46.0% 24.6% 31.1%

Ba c 0.31 0.38 0.42 0.39 -0.10 0.37 0.19 0.13 0.01 0.05 -0.23 0.01 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 5.8% 4.1% 20.1% 2.9% 64.1% 13.5% 26.4% 68.1% 96.4% 78.4% 42.7% 96.7%

Be c 0.22 0.51 -0.12 0.15 -0.08 0.34 0.05 -0.26 -0.16 -0.02 -0.22 -0.01 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 17.8% 0.4% 73.3% 43.2% 69.5% 16.7% 79.4% 41.3% 52.1% 90.8% 44.1% 94.9%

Br c 0.17 0.00 -0.41 0.05 -0.09 -0.10 0.10 0.44 0.32 0.11 -0.31 -0.11 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 29.8% 98.6% 21.6% 79.4% 66.0% 70.0% 56.9% 15.2% 20.1% 56.3% 27.7% 58.7%

Ca c -0.22 0.03 -0.15 0.04 -0.32 -0.11 -0.13 -0.24 -0.37 -0.37 0.20 -0.33 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 17.6% 89.0% 65.0% 81.4% 12.6% 66.3% 44.4% 44.3% 12.6% 5.2% 48.3% 8.8%

Cd c 0.05 0.32 0.20 0.20 -0.21 0.12 -0.17 -0.57 -0.12 -0.42 -0.07 -0.25 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 77.7% 8.4% 55.5% 29.1% 31.3% 64.4% 33.1% 5.3% 64.3% 2.5% 81.6% 19.5%

Cl c 0.18 -0.02 -0.41 0.03 -0.11 -0.04 0.08 0.38 0.29 0.08 -0.31 -0.16 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 27.9% 92.1% 21.2% 87.2% 59.3% 88.0% 65.2% 21.7% 23.6% 69.2% 27.4% 40.5%

Co c 0.07 0.06 -0.02 0.37 -0.61 0.50 -0.29 0.35 -0.25 -0.43 -0.42 -0.14 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 68.3% 75.8% 94.5% 4.1% 0.2% 3.3% 9.3% 27.0% 31.2% 2.1% 13.9% 47.9%

Cond. c 0.26 0.29 0.03 0.06 -0.16 -0.02 0.07 -0.17 -0.15 -0.09 -0.43 -0.10 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 11.8% 11.7% 93.7% 72.9% 45.0% 93.8% 67.7% 58.7% 54.8% 64.6% 12.4% 62.3%

Appendix 235

Appendix 23

As (

g)

Al (

g)

B (g

)

Ba

(g)

Cu

(g)

Fe (g

)

K (g

)

Li (

g)

SiO

2 (g)

Sr (g

)

Zn

(g)

S(V

I) (g

)

Cr c 0.24 0.12 0.40 0.04 -0.20 0.13 0.03 0.02 -0.14 -0.15 0.10 0.04 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 14.5% 53.4% 21.7% 81.6% 35.3% 62.0% 87.5% 94.9% 57.5% 43.3% 72.6% 84.9%

Cu c -0.14 0.08 -0.46 0.35 -0.14 0.62 -0.24 0.20 0.13 -0.38 0.40 0.25 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 40.7% 67.8% 15.4% 5.4% 51.9% 0.6% 17.4% 53.0% 61.4% 4.0% 16.0% 20.8%F c 0.19 -0.04 -0.45 -0.17 0.08 -0.17 0.16 0.45 0.35 0.13 -0.39 0.03 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 24.6% 82.4% 17.0% 35.7% 72.5% 51.0% 35.9% 14.5% 15.7% 49.2% 16.9% 87.3%

Fe(II) c 0.19 0.32 0.43 0.26 -0.12 0.35 0.01 -0.31 -0.13 -0.02 0.43 -0.03 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 25.3% 8.1% 19.0% 16.5% 58.7% 15.1% 95.5% 33.1% 61.8% 91.6% 12.2% 86.5%

Fe(III) c 0.23 0.42 0.10 0.38 -0.29 0.55 -0.04 -0.49 -0.20 -0.17 0.15 -0.08 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 16.1% 2.0% 75.9% 3.4% 16.8% 1.7% 80.6% 10.6% 41.8% 37.4% 60.4% 69.3%

K c 0.09 0.34 -0.31 0.04 -0.17 0.03 0.08 0.17 -0.14 -0.12 -0.24 -0.33 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 57.2% 6.7% 35.5% 84.0% 44.0% 90.0% 64.6% 58.7% 57.0% 54.5% 40.1% 9.0%

Li c 0.20 -0.09 -0.40 0.00 -0.14 -0.17 0.09 0.47 0.37 0.10 -0.31 -0.15 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 21.7% 62.4% 22.3% 98.1% 51.4% 51.0% 62.2% 12.4% 13.5% 60.6% 28.1% 43.6%

Mg c -0.19 0.10 0.51 0.09 -0.15 -0.06 -0.14 -0.22 -0.36 -0.16 0.28 0.00 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 25.0% 61.6% 11.0% 63.7% 48.5% 81.7% 41.8% 48.4% 14.2% 39.5% 33.4% 99.1%

Mn c -0.07 0.19 0.34 0.12 -0.30 0.10 -0.12 -0.21 -0.20 -0.24 0.36 -0.09 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 69.4% 31.7% 31.2% 50.5% 15.5% 68.7% 51.0% 51.3% 42.8% 20.5% 20.8% 65.3%

Mo c 0.20 -0.03 -0.53 -0.02 -0.05 -0.01 -0.05 0.24 0.08 -0.02 -0.35 -0.01 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 21.7% 88.2% 9.6% 92.0% 82.0% 96.4% 79.2% 45.4% 76.3% 90.1% 22.6% 96.5%

Na c 0.11 -0.12 -0.47 -0.16 0.01 -0.30 0.08 0.38 0.33 0.15 -0.29 -0.16 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 50.8% 51.9% 14.2% 38.7% 97.1% 23.3% 66.2% 21.7% 18.2% 42.9% 31.1% 42.3%

Ni c 0.10 0.26 -0.13 0.22 -0.16 0.37 0.04 -0.13 -0.05 -0.32 0.25 -0.16 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 54.3% 16.8% 70.7% 23.0% 46.9% 13.1% 83.8% 68.1% 84.5% 8.7% 39.1% 42.0%

N(V) c 0.03 -0.16 0.31 0.49 -0.54 0.23 -0.35 0.53 0.03 -0.20 0.14 -0.13 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 83.6% 39.3% 35.8% 0.5% 0.6% 35.0% 4.2% 7.9% 90.4% 30.1% 64.2% 51.6%

Pb c 0.18 0.07 0.61 0.14 -0.28 0.12 0.09 -0.15 0.04 -0.14 -0.10 -0.09 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 27.6% 70.4% 4.8% 46.1% 18.8% 63.0% 60.7% 64.1% 87.6% 48.2% 72.6% 63.6%

pH c 0.22 0.31 0.73 0.27 -0.11 0.24 0.03 -0.19 -0.29 -0.08 -0.12 0.12 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 18.3% 9.2% 1.1% 14.1% 61.5% 33.4% 84.5% 55.6% 24.5% 68.6% 67.5% 56.0%

S(-II) c -0.17 -0.34 0.44 -0.10 0.40 -0.29 -0.07 -0.13 0.10 0.07 0.51 0.07 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 30.6% 6.7% 17.4% 59.7% 5.0% 24.4% 70.7% 68.1% 68.3% 72.6% 6.4% 74.0%

236 Appendix

Appendix 23

As (

g)

Al (

g)

B (g

)

Ba

(g)

Cu

(g)

Fe (g

)

K (g

)

Li (

g)

SiO

2 (g)

Sr (g

)

Zn

(g)

S(V

I) (g

)

SiO2 c 0.19 0.48 -0.55 -0.19 0.19 -0.01 0.12 0.01 -0.16 0.05 -0.35 -0.04 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 25.7% 0.8% 7.7% 29.9% 37.9% 95.8% 50.7% 96.6% 51.5% 81.5% 21.5% 85.1%

S(VI) c 0.22 0.27 0.69 0.18 -0.18 -0.10 0.04 -0.48 -0.07 -0.07 -0.12 -0.05 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 19.0% 15.1% 1.9% 34.1% 39.3% 69.3% 82.2% 11.8% 77.9% 72.4% 69.2% 79.9%

Sr c -0.23 0.16 0.25 0.04 -0.10 -0.10 0.01 -0.22 -0.36 -0.35 0.14 -0.18 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 17.3% 40.6% 46.7% 84.3% 64.8% 69.3% 94.3% 49.9% 14.5% 6.7% 63.7% 35.0%

temp c 0.18 -0.24 -0.05 -0.33 0.28 -0.09 0.20 0.30 0.68 0.28 -0.02 0.25 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 28.2% 19.9% 87.3% 6.7% 18.9% 71.7% 25.3% 34.7% 0.2% 13.9% 95.8% 19.5%

TIC c -0.35 -0.36 0.61 -0.24 0.29 -0.71 -0.16 -0.59 -0.21 -0.12 0.37 0.07 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 3.2% 5.4% 4.7% 19.5% 16.5% 0.1% 35.5% 4.2% 39.5% 52.3% 19.1% 73.6%

TOC c -0.05 -0.04 0.78 0.39 -0.17 0.15 -0.07 0.11 0.13 -0.04 0.30 0.07 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 78.1% 83.5% 0.4% 3.2% 43.2% 54.8% 67.2% 73.7% 61.6% 82.5% 29.1% 71.5%V c 0.11 0.11 0.50 0.02 -0.20 0.04 0.09 -0.24 -0.15 -0.15 . -0.22 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 50.7% 55.4% 11.7% 93.5% 35.3% 87.0% 61.4% 45.9% 55.4% 43.3% . 26.5%

Zn c 0.12 0.25 0.01 0.40 -0.33 0.58 -0.06 0.24 -0.14 -0.16 -0.01 -0.13 N 38 30 11 31 24 18 35 12 18 29 14 28 sig 48.0% 18.2% 96.8% 2.7% 11.7% 1.1% 71.1% 45.7% 59.3% 42.1% 97.0% 51.2%

Appendix 237

Appendix 24 Arsenic species distribution modeled with PHREEQC 2.8.03 (Parkhurst and Appelo 1999) using database WATEQ4F (V.2.4 with revised data from Nordstrom and Archer 2003); parameters and species consid-ered were pH, Temp, pe, F, Cl, Br, N(V), S(VI), As(III), As(V), Al, B, Ba, Ca, Cd, Cu, K, Li, Mg, Mn, Na, Ni, Pb, Se, Sr, Zn, Si, S(-II), Fe(II), Fe(II), and C(IV); MMAA and DMAA were added to inorganic As(V) for calculations, concentrations given for As(V) complexes refer to the inorganic As(V) fraction actually determined by on-site species separation; only concentrations > 10-08 mol/L are printed in black

Sampling site

H4A

sO3+

H3A

sO3

H2A

sO3-

HA

sO32-

AsO

33-

H3A

sO4

H2A

sO4-

HA

sO42-

AsO

43-

As 3

S 4(H

S)2-

AsS

(OH

)(H

S)-

YNPGG02-2W 1.9E-11 1.9E-05 5.3E-08 7.1E-16 8.6E-25 1.5E-10 3.5E-07 2.8E-08 2.4E-13 3.3E-12 1.0E-06 YNPGG04-1W 6.6E-10 5.8E-06 1.6E-10 2.1E-20 2.2E-31 1.3E-08 2.3E-07 1.5E-10 1.1E-17 6.4E-14 1.1E-08 YNPGG04-2W 8.7E-10 7.7E-06 2.1E-10 2.7E-20 2.9E-31 1.4E-08 2.5E-07 1.6E-10 1.2E-17 1.4E-13 1.4E-08 YNPGG04-4W 7.7E-10 5.5E-06 1.3E-10 1.4E-20 1.3E-31 1.3E-08 1.8E-07 9.7E-11 6.1E-18 1.2E-13 1.2E-08 YNPGG05-2W 3.5E-10 2.3E-07 2.4E-13 1.1E-24 4.0E-37 6.1E-09 8.1E-09 2.9E-13 8.7E-22 3.4E-15 6.7E-10 YNPGG05-4W 2.8E-10 2.0E-07 1.9E-13 7.1E-25 2.4E-37 1.6E-09 2.5E-09 9.6E-14 2.7E-22 1.0E-15 4.9E-10 YNPGG06-2W 2.0E-12 2.8E-05 1.4E-06 3.7E-13 8.5E-21 5.6E-12 1.7E-07 2.0E-07 3.0E-11 2.1E-12 5.0E-06 YNPGG07-2W 1.3E-11 2.9E-05 2.1E-07 8.1E-15 2.7E-23 6.6E-11 3.0E-07 5.2E-08 1.2E-12 9.8E-11 5.1E-06 YNPGG07-4W 1.4E-11 3.0E-05 2.2E-07 7.9E-15 2.5E-23 3.8E-11 1.7E-07 3.0E-08 6.3E-13 1.6E-11 2.8E-06 YNPGG08-4W 1.8E-08 8.3E-06 8.0E-12 3.6E-23 1.5E-35 3.3E-10 3.3E-10 1.0E-14 3.3E-23 1.0E-22 1.8E-12 YNPHL01-1W 3.6E-10 8.2E-07 6.6E-13 1.8E-24 6.1E-37 3.1E-09 2.3E-08 2.7E-12 1.1E-20 1.2E-12 1.1E-08 YNPHL01-2W 1.3E-09 3.0E-06 3.2E-12 1.2E-23 5.9E-36 5.7E-09 4.0E-08 5.2E-12 2.9E-20 1.1E-10 5.2E-08 YNPHL01-3W 1.4E-09 2.7E-06 2.3E-12 7.1E-24 2.9E-36 9.8E-10 6.1E-09 6.7E-13 3.1E-21 2.2E-09 1.3E-07 YNPHL02-1W 1.5E-13 2.8E-07 2.6E-10 8.5E-19 3.1E-28 2.0E-11 1.1E-07 9.9E-09 3.8E-14 7.0E-14 3.8E-07 YNPHL03-3W 7.8E-10 1.1E-06 2.0E-12 1.3E-23 8.6E-36 1.5E-08 5.0E-08 4.1E-12 2.2E-20 2.3E-25 4.8E-13 YNPLG01-2W 3.9E-13 1.3E-05 1.4E-06 8.2E-13 3.8E-20 1.8E-12 1.1E-07 2.7E-07 8.5E-11 1.2E-12 7.0E-06 YNPLG02-2W 5.5E-15 4.4E-06 1.2E-05 1.7E-10 2.0E-16 8.4E-15 1.3E-08 8.1E-07 6.1E-09 1.0E-19 2.6E-07 YNPLG03-2W 6.5E-14 2.2E-06 2.6E-07 1.5E-13 7.2E-21 1.2E-12 7.9E-08 2.1E-07 6.7E-11 7.2E-12 1.3E-05 YNPNL01-1W 1.5E-10 1.6E-07 6.2E-14 7.8E-26 1.2E-38 9.5E-08 3.3E-07 1.7E-11 3.1E-20 8.8E-21 1.3E-11 YNPNL02-1W 3.8E-09 1.9E-06 1.6E-12 5.5E-24 1.8E-36 1.4E-08 1.7E-08 5.3E-13 1.4E-21 9.1E-12 8.3E-09 YNPNL02-2W 5.9E-09 3.3E-06 4.2E-12 2.4E-23 1.2E-35 1.2E-08 1.4E-08 5.3E-13 1.9E-21 2.8E-09 6.1E-08 YNPNL02-3W 5.3E-10 3.6E-07 3.0E-13 9.7E-25 2.8E-37 7.9E-09 1.2E-08 4.2E-13 1.0E-21 9.5E-14 2.1E-09 YNPNL03-1W 2.4E-10 4.5E-07 1.7E-12 2.6E-23 3.1E-35 1.4E-08 5.0E-08 5.4E-12 4.9E-20 4.1E-16 6.8E-10 YNPNL03-2W 2.1E-10 1.9E-07 4.4E-13 4.6E-24 3.7E-36 4.4E-09 7.3E-09 3.8E-13 2.2E-21 4.7E-16 4.3E-10 YNPNL03-3W 6.3E-10 2.7E-07 3.1E-13 1.7E-24 7.4E-37 2.1E-09 1.8E-09 5.2E-14 1.6E-22 1.5E-15 4.0E-10 YNPNL04-1W 1.3E-10 2.6E-07 3.9E-13 2.0E-24 1.1E-36 7.2E-09 3.6E-08 3.6E-12 1.8E-20 5.2E-13 7.7E-09 YNPNL04-2W 4.1E-10 1.6E-07 8.6E-14 1.9E-25 4.0E-38 2.3E-09 2.0E-09 4.7E-14 8.2E-23 3.4E-25 2.3E-13 YNPNL04-3W 5.2E-10 2.0E-07 1.4E-13 4.0E-25 1.1E-37 3.6E-10 3.0E-10 7.7E-15 1.5E-23 3.5E-15 5.0E-10 YNPNL05-1W 1.0E-08 1.9E-06 7.6E-13 1.5E-24 2.8E-37 9.3E-09 4.0E-09 5.7E-14 8.5E-23 1.0E-13 1.0E-09 YNPNL05-2W 1.2E-08 1.7E-06 7.7E-13 1.9E-24 4.2E-37 5.0E-10 1.7E-10 2.1E-15 3.4E-24 1.1E-15 2.0E-10 YNPRH01-1W 1.1E-08 1.1E-05 1.7E-11 1.1E-22 7.4E-35 7.2E-08 1.6E-07 9.8E-12 5.1E-20 8.4E-12 1.2E-08 YNPRH01-2W 8.4E-09 8.4E-06 1.6E-11 1.4E-22 1.1E-34 9.6E-08 2.1E-07 1.4E-11 8.7E-20 7.8E-13 5.8E-09 YNPRH01-3W 7.6E-09 6.8E-06 1.2E-11 9.9E-23 7.0E-35 4.3E-08 8.3E-08 4.9E-12 2.7E-20 9.7E-11 2.7E-08 YNPRH03-1W 4.7E-08 2.0E-05 6.3E-12 7.9E-24 1.1E-36 7.1E-08 8.7E-08 2.2E-12 3.1E-21 8.3E-20 1.6E-11 YNPRH04-1W 4.7E-08 1.9E-05 7.4E-12 1.2E-23 2.0E-36 3.2E-08 3.6E-08 9.6E-13 1.5E-21 2.6E-15 4.9E-10 YNPRH04-2W 5.2E-08 2.0E-05 7.6E-12 1.2E-23 2.0E-36 9.3E-08 9.9E-08 2.5E-12 4.0E-21 1.3E-17 8.1E-11 YNPRH04-3W 6.4E-08 2.0E-05 6.7E-12 9.6E-24 1.5E-36 1.8E-08 1.5E-08 3.1E-13 4.2E-22 4.0E-13 2.2E-09 YNPRH05-1W 5.4E-08 7.4E-05 1.7E-10 1.9E-21 2.0E-33 1.0E-07 3.3E-07 3.2E-11 2.6E-19 3.2E-06 1.2E-06 YNPRH05-3W 9.7E-08 1.4E-04 6.1E-10 1.4E-20 2.6E-32 9.8E-08 2.8E-07 2.9E-11 3.7E-19 1.7E-06 9.7E-07 YNPRH07-2W 2.9E-09 1.1E-06 1.1E-12 5.2E-24 1.9E-36 7.2E-08 5.2E-08 1.3E-12 3.4E-21 4.2E-15 5.2E-10

238 Appendix

Appendix 24

Sampling site

H4A

sO3+

H3A

sO3

H2A

sO3-

HA

sO32-

AsO

33-

H3A

sO4

H2A

sO4-

HA

sO42-

AsO

43-

As 3

S 4(H

S)2-

AsS

(OH

)(H

S)-

YNPRH08-1W 1.6E-08 3.4E-05 1.6E-10 3.7E-21 8.1E-33 1.2E-07 6.0E-07 1.0E-10 1.7E-18 1.2E-10 5.5E-08 YNPRH08-4W 3.9E-11 3.2E-05 7.8E-08 9.6E-16 1.0E-24 2.3E-10 4.1E-07 2.6E-08 2.0E-13 5.8E-11 2.2E-06 YNPRH09-1W 2.0E-12 3.2E-05 1.3E-06 2.6E-13 4.8E-21 2.8E-12 1.0E-07 1.2E-07 1.7E-11 3.5E-17 1.4E-07 YNPRH10-4W 3.7E-09 3.2E-05 1.0E-09 1.6E-19 2.2E-30 9.7E-09 1.7E-07 1.1E-10 1.0E-17 1.9E-19 1.6E-10 YNPRH11-3W 8.9E-08 3.3E-05 2.2E-11 7.2E-23 2.3E-35 1.0E-07 9.8E-08 2.8E-12 7.5E-21 2.1E-06 4.5E-07 YNPRH13-3W 1.8E-09 3.6E-05 1.8E-09 4.7E-19 1.1E-29 1.5E-08 7.0E-07 1.1E-09 1.9E-16 4.4E-19 3.7E-10 YNPRH14-3W 6.3E-10 3.8E-05 6.9E-09 6.3E-18 5.1E-28 5.0E-09 6.5E-07 3.2E-09 1.8E-15 1.7E-15 1.2E-08 YNPRH15-3W 2.7E-08 2.8E-05 2.1E-11 6.4E-23 2.3E-35 2.5E-07 7.8E-07 5.1E-11 1.9E-19 4.7E-22 5.0E-12 YNPRH16-4W 2.0E-08 3.8E-05 5.4E-11 3.1E-22 2.1E-34 1.1E-07 5.8E-07 6.9E-11 4.6E-19 2.0E-21 1.2E-11 YNPRH17-4W 1.4E-08 1.6E-05 3.7E-11 4.1E-22 4.3E-34 1.1E-07 2.8E-07 2.3E-11 1.8E-19 1.7E-21 8.4E-12 YNPRH18-4W 5.0E-09 4.3E-06 1.1E-11 1.4E-22 1.4E-34 6.9E-08 1.1E-07 6.5E-12 4.2E-20 4.3E-23 2.0E-12 YNPRH19-4W 5.1E-09 2.3E-05 3.8E-10 3.3E-20 2.3E-31 2.8E-08 2.6E-07 9.2E-11 4.4E-18 2.0E-13 1.1E-08 YNPRH20-4W 5.4E-09 2.0E-05 1.4E-10 4.6E-21 1.4E-32 4.4E-08 3.8E-07 1.0E-10 2.5E-18 9.3E-21 3.3E-11 YNPRH21-4W 1.0E-08 1.1E-05 3.0E-11 3.7E-22 4.3E-34 1.7E-07 4.2E-07 3.2E-11 2.8E-19 7.5E-22 6.3E-12 YNPRH22-4W 5.3E-09 4.9E-06 5.8E-12 2.9E-23 1.3E-35 1.4E-07 2.9E-07 1.5E-11 6.1E-20 9.1E-16 5.6E-10 YNPRH24-4W 1.9E-08 2.2E-05 3.3E-11 2.1E-22 1.3E-34 2.2E-07 6.5E-07 5.2E-11 2.9E-19 1.9E-18 8.9E-11 YNPRH25-4W 8.7E-09 2.4E-05 2.3E-10 1.1E-20 4.6E-32 2.2E-08 1.2E-07 2.5E-11 7.0E-19 5.3E-13 1.0E-08 YNPRH26-4W 1.0E-08 1.5E-05 5.3E-11 8.6E-22 1.3E-33 4.5E-08 1.5E-07 1.7E-11 1.8E-19 1.7E-06 1.0E-06 YNPRH27-4W 1.3E-09 3.3E-05 2.8E-09 1.2E-18 4.5E-29 1.2E-08 6.4E-07 1.3E-09 3.3E-16 6.9E-16 5.1E-09 YNPRH28-4W 2.1E-09 1.8E-05 3.6E-10 3.3E-20 2.8E-31 1.7E-08 3.3E-07 2.1E-10 1.2E-17 2.9E-17 8.4E-10

Appendix 239

Appendix 25 Saturation indices for As minerals modeled with PHREEQC 2.8.03 (Parkhurst and Appelo 1999) using database WATEQ4F (V.2.4 with revised data from Nordstrom and Archer 2003); for parameters and spe-cies considered see Appendix 24; supersaturated phases marked in bold; other As minerals listed in the WATEQ4F database and not displayed in the table below (As2O5; AlAsO4:2H2O; Ca3(AsO4)2:4H2O; Cu3(AsO4)2:6H2O; Mn3(AsO4):8H2O; Ni3(AsO4)2:8H2O; Pb3(AsO4)2; Zn3(AsO4)2:2.5H2O) are all undersaturated

Sampling site

Scor

odite

Ars

enol

ite

Cla

udet

ite

Orp

imen

t

As 2

S 3(a

m)

Rea

lgar

Sampling site

Scor

odite

Ars

enol

ite

Cla

udet

ite

Orp

imen

t

As 2

S 3(a

m)

Rea

lgar

YNPGG02-2W -6.3 -8.8 -8.8 -7.9 -8.8 -6.0 YNPRH01-1W -4.9 -9.2 -9.2 -4.6 -5.7 -6.8 YNPGG04-1W -7.0 -10.0 -10.0 -8.8 -9.6 -6.4 YNPRH01-2W -5.1 -9.5 -9.5 -5.9 -6.8 -9.4 YNPGG04-2W -7.0 -9.7 -9.7 -8.6 -9.4 -6.3 YNPRH01-3W -5.9 -9.7 -9.7 -5.1 -6.0 -7.9 YNPGG04-4W -6.7 -10.0 -10.0 -8.8 -9.7 -10.0 YNPRH03-1W -4.8 -8.3 -8.3 -5.0 -6.3 -13.7YNPGG05-2W -8.8 -12.6 -12.6 -7.4 -8.4 -8.7 YNPRH04-1W -4.6 -8.4 -8.4 -3.6 -4.8 -13.3YNPGG05-4W -8.0 -12.6 -12.6 -6.8 -7.9 -7.3 YNPRH04-2W -5.1 -8.3 -8.4 -4.8 -6.0 -12.3YNPGG06-2W -8.1 -8.6 -8.6 -9.6 -10.4 -6.5 YNPRH04-3W -5.4 -8.4 -8.4 -3.1 -4.2 -6.4 YNPGG07-2W -7.2 -8.6 -8.6 -8.3 -9.1 -7.1 YNPRH05-1W -3.9 -7.5 -7.5 -1.6 -2.6 -5.7 YNPGG07-4W -7.3 -8.6 -8.6 -8.6 -9.4 -7.2 YNPRH05-3W -4.5 -7.3 -7.3 -4.2 -5.0 -7.4 YNPGG08-4W -6.9 -9.5 -9.5 -11.0 -12.0 -8.3 YNPRH07-2W -6.9 -11.4 -11.4 -8.8 -9.6 -7.3 YNPHL01-1W -4.5 -10.7 -10.7 0.4 -1.0 -7.7 YNPRH08-1W -6.5 -8.2 -8.2 -4.8 -5.8 -5.4 YNPHL01-2W -5.0 -9.7 -9.7 0.6 -0.8 -7.9 YNPRH08-4W -6.2 -8.5 -8.5 -7.6 -8.5 -8.3 YNPHL01-3W -5.1 -9.8 -9.8 1.6 0.2 -3.1 YNPRH09-1W -7.6 -8.4 -8.4 -10.7 -11.6 -9.6 YNPHL02-1W -4.8 -11.8 -11.9 -3.5 -4.9 -7.8 YNPRH10-4W -5.5 -8.6 -8.6 -11.8 -12.6 -4.2 YNPHL03-3W -5.9 -11.0 -11.0 -10.9 -12.0 -6.5 YNPRH11-3W -6.1 -8.2 -8.2 -1.5 -2.5 -3.9 YNPLG01-2W -8.9 -9.4 -9.3 -10.3 -11.1 -6.6 YNPRH13-3W -5.8 -8.3 -8.3 -10.2 -11.1 -12.7YNPLG02-2W -12.0 -10.3 -10.3 -14.7 -15.5 -8.5 YNPRH14-3W -5.8 -8.3 -8.3 -9.2 -10.1 -10.7YNPLG03-2W -8.7 -10.9 -10.9 -10.2 -11.0 -5.8 YNPRH15-3W -5.4 -7.9 -7.9 -6.1 -7.4 2.4 YNPNL01-1W -5.6 -12.1 -12.2 -4.0 -5.4 -7.4 YNPRH16-4W -5.9 -7.7 -7.7 -6.0 -7.3 1.8 YNPNL02-1W -7.2 -10.7 -10.7 -4.8 -5.8 -9.0 YNPRH17-4W -5.2 -8.9 -8.9 -10.3 -11.2 -9.9 YNPNL02-2W -8.6 -10.3 -10.3 -4.8 -5.7 -5.0 YNPRH18-4W -6.4 -10.3 -10.2 -13.1 -13.9 -8.9 YNPNL02-3W -7.9 -12.0 -12.0 -5.5 -6.5 -6.1 YNPRH19-4W -6.6 -8.8 -8.8 -8.7 -9.5 -7.5 YNPNL03-1W -5.8 -12.0 -12.0 -8.7 -9.6 -13.1 YNPRH20-4W -5.8 -8.7 -8.7 -9.7 -10.6 -14.1YNPNL03-2W -6.9 -12.9 -12.9 -9.9 -10.7 -8.8 YNPRH21-4W -5.9 -9.3 -9.3 -11.0 -12.0 -8.7 YNPNL03-3W -7.2 -12.6 -12.6 -9.3 -10.2 -7.4 YNPRH22-4W -5.2 -9.8 -9.8 -6.0 -7.1 -5.2 YNPNL04-1W -6.7 -12.1 -12.1 -3.3 -4.5 -4.7 YNPRH24-4W -5.8 -8.4 -8.4 -6.6 -7.7 -8.6 YNPNL04-2W -8.4 -12.8 -12.8 -11.5 -12.5 -8.7 YNPRH25-4W -5.3 -8.8 -8.8 -8.3 -9.1 -6.9 YNPNL04-3W -7.8 -12.7 -12.7 -7.4 -8.3 -8.8 YNPRH26-4W -5.9 -9.0 -9.0 -3.2 -4.1 -2.4 YNPNL05-1W -6.8 -10.8 -10.8 -6.5 -7.4 -11.9 YNPRH27-4W -5.8 -8.5 -8.5 -9.8 -10.6 -9.7 YNPNL05-2W -7.2 -11.0 -11.0 -8.7 -9.6 -12.2 YNPRH28-4W -5.6 -8.9 -8.9 -9.0 -9.9 -8.8

240 Appendix

Appendix 26 Thermodynamic data for gaseous As species from Spycher and Reed (1989) and Pokrovski et al. (2002) Spycher and Reed (1989): As(g) As + 0.375SO4-2 + 0.375H+ + 1.5H2O = H3AsO3 + 0.375HS- log_k 45.78 As2(g) As2 + 0.75SO4-2 + 0.75H+ + 3H2O = 2H3AsO3 + 0.75HS- log_k 24.89 As3(g) As3 + 1.125SO4-2 + 1.125H+ + 4.5H2O = 3H3AsO3 + 1.125HS- log_k 35.61 As4(g) As4 + 1.5SO4-2 + 1.5H+ + 6H2O = 4H3AsO3 + 1.5HS- log_k 18.22 AsCl3(g) AsCl3 + 3H2O = H3AsO3 + 3Cl- + 3H+ log_k 12.97 AsF3(g) AsF3 + 3H2O = 3F- + H3AsO3 + 3H+ log_k 0.58 AsF5(g) AsF5 + 0.25HS- + 4H2O = H3AsO3 + 5.25H+ + 0.25SO4-2 + 5F- log_k 20.97 AsH3(g) AsH3 + 0.75H+ + 0.75SO4-2 = H3AsO3 + 0.75HS- log_k 24.86 As4O6(g) As4O6 + 6H2O = 4H3AsO3 log_k 8.01 AsS(g) AsS + 0.125SO4-2 + 2.5H2O = H3AsO3 + 1.125HS- + 0.875H+ log_k 13.65 As4S4(g) As4S4 + 0.5SO4-2 + 10H2O = 4H3AsO3 + 4.5HS- + 3.5H+ log_k -51.05 Pokrovski et al. (2002): As(OH)3(g) H3AsO3 = H3AsO3 log_k 11.3

Appendix 241

Appendix 27 Partial pressure of gaseous As species considering the thermodynamic data from Spycher and Reed (1989) and Pokrovski et al. (2002; Appendix 26) in PHREEQC; for parameters and dissolved species considered see Appendix 24

Sampling site A

s(g)

As 2

(g)

As 3

(g)

As 4

(g)

AsC

l 3(g)

AsF

3(g)

AsF

5(g)

AsH

3(g)

AsS

(g)

As 4

S 4(g

)

As 4

O6(

g)

As(

OH

) 3(g

)

YNPGG02-2W 1.8E-48 1.5E-31 3.2E-46 8.4E-33 1.1E-39 6.2E-32 1.9E-71 8.4E-27 2.0E-29 7.3E-16 1.4E-25 9.8E-15










YNPHL01-1W 1.6E-51 1.3E-37 2.4E-55 5.8E-45 1.1E-34 3.7E-36 1.9E-76 1.7E-31 2.7E-30 2.5E-19 4.4E-31 4.1E-16





YNPLG01-2W 2.3E-47 2.4E-29 6.4E-43 2.2E-28 1.0E-44 7.4E-35 4.8E-77 2.0E-24 2.1E-29 8.6E-16 2.6E-26 6.4E-15



YNPNL01-1W 8.0E-53 3.0E-40 2.7E-59 3.2E-50 1.0E-36 1.0E-33 1.7E-71 2.0E-33 3.8E-32 9.5E-27 6.5E-34 8.0E-17












YNPRH01-1W 9.5E-51 4.2E-36 4.6E-53 6.5E-42 1.7E-31 1.3E-32 4.7E-71 4.3E-31 1.2E-29 9.1E-17 1.2E-26 5.3E-15













242 Appendix

Sampling site

As(

g)

As 2

(g)

As 3

(g)

As 4

(g)

AsC

l 3(g)

AsF

3(g)

AsF

5(g)

AsH

3(g)

AsS

(g)

As 4

S 4(g

)

As 4

O6(

g)

As(

OH

) 3(g

)


















Documents

VOLATILE ARSENIC IN AQUATIC ENVIRONMENTS€¦ · troduced until the early 1970s with the invention of the hydride generation technique. Arsenic bio-methylation pathways in different